Clinical Trials that May Change What You Were Taught in Pharmacy School: Part 2
Steph’s Note: Today we’re bringing back a new crowd favorite - Christian Gonzalez-Hernandez. You may remember him from his awesomely comprehensive diabetes clinical trial review a couple of months ago - and if you don’t, then definitely visit his tl;dr debut here. You won’t regret it!
Because Christian is apparently being ridiculously productive during this quarantine, he has already busted out his next round of clinical trial reviews for y’all. And this time he’s gone to infectious diseases. So have fun learning about some brand spanking new literature!
And FYI, if you need some help keeping antibiotics straight, check out this FREE cheat sheet!
Infectious diseases (ID) has been one of my areas of interest ever since I came across my first case of vancomycin dosing. Perhaps this is because it is one area in which pharmacists can have a substantial amount of involvement - from antimicrobial stewardship to IV to PO conversions and PK monitoring to recommendations about duration of therapy.
That’s right. We’re talking to you, (future and current) pharmacists! (Image)
Pharmacists definitely make an impact.
As of late, ID trials are mostly focused on improving outcomes, assessing the non-inferiority of shorter courses, and assessing ways to monitor the efficacy of antimicrobials using their respective PK/PD properties.
So let’s dive into some ID updates.
Without a doubt, sepsis and septic shock represent two of the most important topics when talking about inpatient medicine and infectious diseases. Sepsis has a significant mortality rate and is responsible for about one-third of all hospital deaths. This makes sepsis the leading cause of death in US hospitals.
Whoa.
No wonder it’s imperative that new and improved treatment options are constantly explored and evaluated!
So What is Sepsis?
As a refresher, sepsis is an unregulated response by the host to infection. It may include physiologic, metabolic, and biochemical abnormalities, which can lead to organ dysfunction and even failure.
Before 2016, the SIRS (Systemic Inflamatory Response Syndrome) criteria were used to identify sepsis. Then in 2016, the Third International Consensus Definitions of Sepsis and Septic Shock were released. This group basically said the SIRS criteria weren’t helpful for identifying sepsis. Here’s their rationale for this determination:
Increased temperature, heart rate, respiratory rate, and/or white cell blood count do not necessarily mean the host is having an unregulated, life-threatening response.
SIRS criteria can be met in patients who aren’t even fighting an infection!
Sometimes SIRS criteria are met just from an adaptive response to stress rather than organ dysfunction or systemic inflammation.
On the other hand, many patients who do actually have sepsis may not exhibit the 2 or more SIRS criteria required to meet that sepsis definition.
So basically, sensitivity and specificity of the SIRS criteria for identifying patients with sepsis are pretty crappy.
So the search for a method to assess sepsis and organ dysfunction continued, and it led to the SOFA score.
The current predominant score to identify sepsis is the Sequential Organ Failure Assessment (SOFA). The SOFA score uses lab parameters as markers of organ function. Check out the parameters and assessments below:
The SOFA score parameters in a nutshell. (Image)
NOTE: The higher the SOFA score, the higher the probability of death.
The 2016 panel recommended that a change of at least 2 in baseline SOFA score may be appropriate to identify organ dysfunction or a dysregulated response. It is important to note a couple things when using the SOFA score:
The baseline SOFA score should be assumed to be 0 unless the patient is known to have acute or chronic organ dysfunction before the onset of infection.
A score of 2 correlates to a mortality risk of 10% (yowsers!)
Many of the variables used may require lab testing, and so may not be immediately available.
Pay special attention to item number 3.
Sepsis is an emergency where every second counts. Therefore a screening tool that can be performed quickly is beneficial for identifying patients who may potentially be septic.
Enter qSOFA (q stands for quick). The qSOFA, which is a validated clinical model developed with multivariable logistic regression, identifies potentially septic patients if they meet 2 of the following 3 criteria:
GCS of 13 or less, aka altered mental status,
Systolic blood pressure of 100 mmHg or less, and/or
Respiratory rate greater than 22 breaths/min.
A couple important notes about both the SOFA and qSOFA scores. The SOFA score is really best applied to assess ICU patients’ mortality risk. It’s not meant to assess effectiveness of treatment or monitor improvement over time. The qSOFA score is most useful for patients outside of the ICU as the SOFA correlated better with organ dysfunction and mortality than qSOFA in ICU patients.
Now that we’ve discussed what currently constitutes sepsis and how to identify it, let’s discuss how to tackle the problem.
Management of Sepsis
In 2018, the updated Surviving Sepsis Campaign was released. Along with this, there was an updated 1-hour version of the “sepsis bundle”. To properly treat sepsis, there is a series of actions that need to take place within 1 hour of the recognition of sepsis and septic shock (more on the definition of septic shock later), including:
Measurement of lactate (if lactate is > 2 mmol/L, recheck lactate level)
Obtain blood cultures before administration of antibiotics
Administration of broad spectrum antibiotics
Administration of 30 mL/kg of crystalloid fluids for hypotension or elevated lactate (>4 mmol/L)
Administration of vasopressors to maintain MAP above 65 mmHg, if the patient remains hypotensive after fluid resuscitation.
Points 1, 2, and 4 are more straightforward and do not require a ton of explanation. But I would like to expand on points 3 and 5.
How it feels sometimes when recommending empiric therapy… (Image)
Selection of Antimicrobials in Sepsis
The selection of antimicrobials will be dependent on multiple factors, including: suspected infection site, local susceptibility patterns, patient factors (e.g., allergies, previous infections, and risk factors for resistance), and current resources (availability of antimicrobials, shortages). There are probably multiple other factors that are relevant when choosing an antimicrobial regimen, but for the sake of simplicity, those cover most bases.
These factors are more important when we are talking about empiric treatment with antimicrobials, i.e., treatment that is initiated before any definitive culture results. Think about empiric therapy as an ‘educated guess’ about what kind of coverage is needed. Experience, risk factors, and current data of the most likely pathogens are used to guide the initial treatment choices.
In sepsis, the identification of a likely source of infection is important to determine which antibiotics would be the best options. For empiric treatment, the Surviving Sepsis Guidelines recommend the administration of broad-spectrum antimicrobials to cover all most likely pathogens.
In most instances, this is how I tackle this:
Depending on the suspected source of infection, we may be trying to cover both gram-positive and gram-negative organisms.
Nowadays the prevalence of MRSA is significant, and coverage with a MRSA-active agent should be considered.
Vancomycin is the standard workhorse for empiric MRSA coverage (although it still should be a thoughtful decision rather than a knee jerk reaction - more on vancomycin later).
Given its bacteriostatic mechanism of action, there is some controversy regarding the use of linezolid for bacteremia even though technically it “covers” MRSA. Depending on the suspected source of infection, linezolid may not be the best first choice for a septic patient.
Daptomycin is another possibility for MRSA coverage if vancomycin is not appropriate; however, remember that it is inactivated by surfactant in the lungs and so should not be used for sepsis of suspected pulmonary origin!
Gram-negative coverage tends to be more complicated since the incidence of resistant organisms varies by region. Heck, it even varies from hospital to hospital and sometimes even unit to unit! Therefore, having a copy of your local antibiogram nearby is always useful. (If you need a refresher on how to interpret these tools, check this out.) That being said, some options may include:
Cefepime, if ESBL prevalence is low in your area
A carbapenem if ESBL prevalence is significant in your area (usually resistance rates over 10-15%) or your patient has a history of these resistant organisms
Aztreonam for patients that have true allergies to beta lactams
Fluoroquinolones may be options if:
The patient is allergic to beta-lactams
Susceptibility patterns in your area show low levels of resistance (less than 10-20% resistance)
Cost or availability of other options is a significant barrier
Once culture and susceptibility data become available, it is appropriate to step down antimicrobial therapy. What if nothing grows (yes, sometimes this happens!)? The guidelines also suggest stepping down if clinical improvement is noted.
(This gets complicated if there is no culture or susceptibility data because how do you know what medication to step down to…more empiric decisions. And monitoring for clinical worsening after modifying the regimen.)
Furthermore, even though it is a weak recommendation, the Surviving Sepsis Guidelines recommend that combination antibiotic therapy may be given in patients with septic shock.
What is Septic Shock?
Septic shock is usually defined as hypotension that requires the administration of vasopressors to maintain a MAP of at least 65 mmHg. Vasopressors are the drugs of choice to help with this when fluids are not sufficient.
The waters get so deep and dark when it comes to refractory septic shock treatment that you just might run into this guy. (Image)
Now, when one or more pressors is not doing the trick, the waters get deeper and we enter some sort of gray area…that will get darker and deeper with every line of treatment we attempt to cross. We may get into refractory septic shock.
In these patients, there are several approaches to treatment, but concerns and controversy surround most of these. This is mostly because there isn’t a robust amount of quality evidence to support their uses. Among these options we may find:
Albumin: Usually this colloid is tried in patients who require large amounts of crystalloids in order to maintain intravascular fluid. But crystalloids are generally preferred because there was no difference in outcomes between the 2 types of resuscitation. Plus crystalloids are far more economical!
Corticosteroids: Most times, when you talk about corticosteroids as a treatment option in sepsis, controversy abounds. The 2018 ADRENAL trial found no difference in terms of mortality but did find differences in discharge from the ICU and cessation of mechanical ventilation.
It is worth noting that the 2018 APROCCHSS trial found a difference in terms of mortality at 90 days when hydrocortisone and fludrocortisone were given in combination. The Surviving Sepsis Guidelines only recommend the use of 200 mg hydrocortisone/day in those patients that remain hemodynamically unstable despite the use of fluids and vasopressors.
Lower the MAP target: The rationale behind this one is that excessive dependence on vasopressors to achieve a MAP target of 65 mmHg may reduce blood flow in vasoconstricted vascular beds. The 65 trial published earlier in 2020 assessed whether or not a MAP target of 60 mmHg would improve 90-day mortality. Even though numerically there was a reduction in 90-day mortality in the intervention group, there was not a significant difference.
This would be a good journal club article if you are on a critical care rotation (wink, wink).
Replacement of vitamin C and thiamine: Enter the Vitamins trial.
The VITAMINS Trial
The VITAMINS trial was one of the first controversial trials released in 2020.
The use of vitamin C has been theorized to benefit patients with septic shock due to several reasons. Some of those reasons are:
Vitamin C has some anti-inflammatory and antioxidant properties,
It plays a role in preserving endothelial function, supporting immune function, and may even serve as a co-factor for enzymes needed for the synthesis of vasopressors, and
In critically ill patients, there is some data supporting that thiamine and vitamin C levels are low.
In 2014 there was a small, phase I study with only 24 participants. Patients were randomized 1:1:1 to receive low dose vitamin C (50 mg/kg/day), high-dose vitamin C (200 mg/kg/day), or placebo. Reductions in SOFA score were seen in the groups who received vitamin C.
Fast forward to 2017, a retrospective before-after study assessing hospital survival was performed in patients with severe sepsis and a procalcitonin level >2 ng/mL. Forty-seven ICU patients received IV vitamin C (1.5 gm every 6 hours for 4 days or until ICU discharge), hydrocortisone (50 mg IV every 6 hours for 7 days or until ICU discharge followed by a 3-day taper), and thiamine (200 mg every 12 hours for 4 days or until ICU discharge). These patients were compared to a prior control group of 47 patients who did not receive the above standard regimen. Significant differences were seen in mortality, duration of vasopressors, and change in SOFA score at 72 hours.
In the 2020 VITAMINS trial, 216 patients in 10 sites from Australia, New Zealand, and Brazil were randomized in a 1:1 ratio to either the same vitamin cocktail from the 2017 study or standard of care with hydrocortisone alone. It is worth noting that administration of thiamine was allowed in the control group at the discretion of the attending ICU clinician. This was the first study that compared the combination of vitamin C, thiamine, and hydrocortisone vs hydrocortisone alone. Patients needed to be enrolled in the study within 24 hours of a septic shock diagnosis according to the Sepsis-3/2018 Surviving Sepsis Guidelines criteria (as discussed above).
The primary outcome for this study was length of time alive and free of vasopressors up until day 7, even if patients ultimately died. In simple terms, if vasopressors were weaned and the patient died thereafter, those hours were free of pressers were still counted towards the primary outcome. Vasopressor discontinuation was defined as not receiving vasopressors for a 4 hour period. There were multiple secondary outcomes measured, including:
28- and 90-day mortality rates
ICU and hospital mortality rates
28-day cumulative vasopressor-free days
28-day cumulative mechanical ventilation-free days
28-day renal replacement therapy–free days
28-day ICU free-days
Change in SOFA score at day 3
Hospital length of stay
No difference was seen between both groups in any of the primary or secondary outcomes, except for change in SOFA score at 72 hours. Also, there was no difference in vasopressor re-dependence by day 7. This suggests that the hydrocortisone, thiamine, and vitamin C combination did not provide any additional benefit compared against standard of care plus hydrocortisone.
But before we rush and make a decision, we need to look into the details and see what differs between this study and previous studies. The VITAMINS study assessed differences up to a longer duration than the previous observational study conducted in 2017. In the older study, the effects of treatment were assessed for up to 4 days of therapy.
One more interesting piece of information about the 2017 study… That control group received “standard of care.” Bear in mind, though, that while those patients didn’t receive vitamin C or thiamine, it was left to the discretion of the clinician whether or not hydrocortisone would be administered.
Turns out 28 of the 47 control group patients (about 60%) actually did receive hydrocortisone! So perhaps the benefits observed in the intervention group were indeed due to vitamin C or thiamine given hydrocortisone was used in both groups? Hard to say for sure given this was a retrospective study, but certainly food for thought.
Moving ahead, looking at the interventions…
The VITAMINS trial compared the triple combination cocktail vs. hydrocortisone alone in a randomized fashion, allowing us to get a better idea of whether any observed benefits were due to hydrocortisone or thiamine and vitamin C. Even though the authors noted that thiamine administration in the hydrocortisone group was allowed at the discretion of the ICU clinicians, it was a pretty small percentage of the control group that got both steroids and thiamine. But because of the crossover of therapies, it sure would have been nice to have had an analysis of the benefit of each component of the triple cocktail…wouldn’t it?? Sigh.
Also the VITAMINS trial was open-label and unblinded. Did someone say possible confounders?
You may remember from my previous post that I mentioned there is important information from secondary outcomes in clinical trials. This is no exception.
Even though there were no significant differences in most secondary outcomes of the VITAMINS trial, the study was not powered to detect differences in secondary outcomes - only the primary outcome. As a matter of fact, the sample size in this trial was not really that big (N=216).
In my opinion, this is why they decided to use hours as the primary outcome instead of the patient survival rate! With the hours primary outcome, It’s easier to detect differences with a lesser number of participants, just like the use of composite endpoints allows investigators to observe a certain number of events to reach statistical power.
This study highlights the importance of performing randomized controlled trials. It also emphasizes how randomization helps us figure out if differences seen in observational studies are due to the intervention itself or just differences between groups (selection bias).
When your inbox blows up with COVID literature updates every day. (Image)
(This is especially important to remember in these current times of COVID19, when lots of articles are being published every day - even many articles that have not yet been peer-reviewed or that have a low quality of evidence. Sometimes it may take multiple studies to have a sufficient amount of data to adopt novel treatment strategies.)
On the note of randomized controlled trials (RCTs), there have been some RCTs assessing the utility of just vitamin C in sepsis, but most of them have reported no benefit in mortality. Only the 2019 CITRIS-ALI study and a small phase I trial with 24 participants reported any sort of mortality benefit.
Currently, according to clinicaltrials.org, there are over 30 trials attempting to study vitamin C in sepsis/septic shock! Eighteen are investigating triple therapy (vitamin C, thiamine, hydrocortisone), 12 are testing vitamin C alone, and 3 are looking into thiamine and vitamin C. Hopefully this means that in the coming months/years, we will have some more data to make a more certain recommendation for the role of vitamin C in sepsis/septic shock.
Until then:
There is no data to adopt vitamin C as a universal measure or as the standard of care for sepsis treatment.
That being said, adverse effects have not been problematic when studying vitamin C combinations. Therefore, in patients who have refractory septic shock and are not improving despite the use of other therapies, it may be a consideration - knowing the amount of evidence behind this is not robust.
The Camera-2 Trial
The workhorse for MRSA treatment is good old vancomycin. But of course, no drug is perfect. And vancomycin has plenty of shortcomings, such as variable tissue penetration, prolonged time to effect, and the need for close monitoring.
In the past couple of years, there has been increasing discussion about vancomycin leading to poorer outcomes compared with anti-staphylococcal beta-lactams when combating MSSA infections. This led to the idea that adding an anti-staphylococcal beta-lactam to MRSA-active agents may improve outcomes, and this has even been supported by in vitro data.
Until recently, there were only two small clinical trials evaluating the use of combination therapy for Staphylococcus aureus bacteremia. A small trial involving 60 patients compared vancomycin monotherapy vs. vancomycin plus flucloxacillin with a primary outcome of duration of bacteremia, but it found no mortality difference. Another trial studied the combination of daptomycin and ceftaroline vs. daptomycin monotherapy, which did find a difference in terms of mortality.
So one nay and one yay. Hmm. What to do.
The 2020 CAMERA-2 trial was a multicenter, open-label, randomized trial, in which 352 bacteremic patients were randomized in a 1:1 ratio to receive standard therapy of either vancomycin (adjusted to maintain trough levels of 15-20 mcg/mL) or daptomycin (6-10 mg/kg at the discretion of the provider) or combination therapy with one of the above PLUS an anti-staphylococcal beta-lactam.
The anti-staphylococcal beta lactams used were flucloxacillin 2 g every 6 hours in Australia/New Zealand or cloxacillin 2 g every 6 hours in Singapore/Israel. Beta-lactam duration was 7 days, although the vancomycin and daptomycin courses could be longer.
The study’s primary outcome was a composite that consisted of the following four components:
All-cause mortality
Persistent bacteremia at day 5
Microbiological relapse (defined as a positive culture at least 72 hours after having a negative culture)
Microbiological treatment failure (positive culture 14 days after randomization)
As with many other clinical trials, the secondary outcomes included the components of the primary composite outcome:
All-cause mortality at 14, 42, and 90 days
Persistent bacteremia at days 2 and 5
Acute kidney injury (AKI), defined as stage 1 or higher using modified RIFLE criteria
Microbiological relapse
Microbiological treatment failure
Duration of intravenous antibiotic treatment
This study demonstrated no significant difference in the primary outcome (35% in the combination group vs. 39% in the standard therapy group; P=0.42).
This composite endpoint was a meaningful one, where both clinical markers (resolution of bacteremia) and patient outcomes (mortality) were included. I think that endpoint provided us with a pretty good idea of whether the combination treatment would be beneficial or not. It is a great example of keeping an eye on the patient and not just the number.
Now, every time there is a composite endpoint, it is imperative to look at the individual components of that endpoint. In this case, when we look at mortality, there was an increase in mortality in the combination treatment group, even if it wasn’t statistically significant. At 90 days, 35 patients died in the combination group compared to 28 patients in the control group.
When looking at another component of that composite endpoint, which is persistence of bacteremia at day 5, we also see some differences: 19 patients demonstrated Staph growth in the combination group vs. 35 patients in the standard therapy group (P=0.02).
AKI according to the RIFLE criteria was more common in the combination group than in the standard therapy group. When using the KDIGO criteria in a post-hoc analysis, this did not change - 34 patients in the combination group had AKI vs. only 13 patients in the standard therapy group.
Duration of IV antibiotics was also longer in the combination group, resulting in an overall longer exposure to IV antibiotics.
Now, before delivering a verdict, we need to look into a few other details from this study…
How all those other antibiotic combos feel when we try to group them in with the study drug results from CAMERA-2. (Image)
Participants could technically receive either vancomycin or daptomycin. But the truth is that the study mostly focused on vancomycin and the anti-staphylococcal penicillins. Only 27 patients received cefazolin, while 111 patients received an anti-staphylococcal penicillin. Furthermore, over 90% of patients in this study received vancomycin, with only 4% of patients receiving at least one dose of daptomycin. This is meaningful, because we cannot generalize the study results to other medications or combinations (e.g., cefazolin + vancomycin, cefazolin + daptomycin, or anti-staphylococcal penicillins + daptomycin).
It is also worth noting that vancomycin dosing was guided using trough levels, which is not a sin per se... That being said, trough levels in this study were at the upper end of the 15-20 mcg/mL range, with an average value of 20.1 mcg/mL on day 3 in the combination group and 19.2 mcg/mL in the standard therapy group.
Trough based dosing can understimate the real AUC by about 23%, meaning that trough values at the upper end might lead to more acute kidney failure. As a matter of fact, the new 2020 IDSA vancomycin guidelines recommend AUC-based dosing because it can lead to more efficient target attainment than trough-based dosing.
So perhaps the vancomycin dosing strategy could have influenced the rates of AKI noted in this study.
Another important piece that could have influenced AKI rates is how infrequently cefazolin was used in this study. Only one of the 27 patients who received cefazolin suffered from AKI. Both in this study and in retrospective data from other studies, AKI is less common with cefazolin compared to rates with anti-staphylococcal penicillins. With this being said, clinical trials assessing combination treatment for MRSA infections using daptomycin and/or cefazolin may be warranted in the future and might be interesting to see.
This may be a prime example of the importance of balancing risks and benefits. Even though there was improvement in the resolution of the bacteremia at day-5 with combination therapy, this did not translate into a mortality benefit. Thankfully the composite endpoint was able to detect that.
Another plus about this study is that it used combination treatment from the get-go rather than studying it as salvage therapy for MRSA bacteremia like most previous trials.
So, what’s my final take on CAMERA-2?
Based the fact that mortality rates did not improve and AKI rates were higher, combination vancomycin + anti-staphylococcal penicillin should not be the first choice for empiric MRSA bacteremia therapy.
The DANCE Trial
Now, stepping away from bacteremia and sepsis (but staying within the ID realm), we will discuss another hot topic within the ID community. This one goes out to all the fans of antimicrobial stewardship.
I bet many of you have heard the phrase ‘less is more’ in recent years, and this is often one of my favorite pharmacotherapy approaches - especially when it comes to antibiotics. Prolonged exposure to antimicrobials is one of the driving factors for increased rates of antibiotic resistance. Plus it is associated with increased adverse effects.
Hence, we really need to find ways to avoid unnecessary exposure to these medications.
The “less is more” movement hopes to shorten antimicrobial regimens without affecting outcomes. So far there have been studies supporting shorter courses of therapy in community acquired pneumonia (CAP), hospital acquired pneumonia (HAP),and osteomyelitis, among others which can be seen in the table below:
Before immersing ourselves into this next trial, let’s review some basic concepts and clinical pearls regarding cellulitis. First, we need to classify whether the infection is purulent or non-purulent. In combination with signs and symptoms, this will guide our choice of empiric treatment. According to the 2014 IDSA cellulitis guidelines:
(Image)
For mild cases of cellulitis (no systemic signs and symptoms like fever or chills), the IDSA guidelines recommend the use of an antimicrobial with streptococcal coverage such as: cephalexin 500 mg PO every 6 hours, penicillin 250-500 mg PO every 6 hours, or clindamycin 300-450 mg PO every 6 hours - but beware clindamycin resistance!
In patients with moderate disease (some low-grade fevers), coverage should be broadened to cover Staphylococcus aureus. Helpful antibiotics include: cephalexin 500 mg PO every 6 hours, dicloxacillin 500 mg PO every 6 hours, and cefazolin 1 g IV every 8 hours (if IV is warranted).
Those patients that have risk factors associated with MRSA infections such as: cellulitis is associated with trauma, history of MRSA infections, colonization of MRSA elsewhere, or history of IV drug use need to have coverage against MRSA with clindamycin 300-450 mg PO every 6 hours, doxycycline 100 mg PO twice daily, sulfamethoxazole/trimethoprim (Bactrim) 1-2 DS tablets PO twice daily.
Immunocompromised patients or those patients with severe infection, which is defined as failure of oral antibiotics or having one of the following: temperature >38*C, tachycardia (>90 beats per minute), tachypnea (respiratory rates > 24 breaths per minute, or leukocytosis (WBC >12,000 cells/uL) or leukopenia (WBC <400 cells/uL). These patients need IV antibiotics with MRSA and streptococcal coverage.
Options include: vancomycin dosed according to AUC:MIC > 400 or trough levels between 10-15 mcg/mL, linezolid 600 mg IV every 12 hours, and daptomycin 4 mg/kg every 24 hrs. These patients warrant hospital admission.
In patients with purulent cellulitis, incision and debridement (I&D) and MRSA coverage is warranted. Usually what defines whether someone will be treated with PO antibiotics in the outpatient setting vs. with inpatient IV antibiotics is the presence of systemic signs and symptoms like the ones mentioned in the previous paragraph. The current IDSA recommendation for duration of therapy in cellulitis is 5 days if improvement is reported.
But here’s the truth.
Five days is rarely the case because, even in milder cases, clinicians are afraid shorter courses of treatment may lead to recurrence of the infection. So should we trust the guideline-directed shorter course or not?
The 2020 DANCE trial was a multicenter, double-blind, placebo-controlled, randomized trial of 149 Dutch participants. Cellulitis was defined as warmth, erythema, and induration of the skin and/or subcutaneous tissue, with or without pain, and included erysipelas. It is important to note that this definition left out other important markers of infection, such as fever, leukocytosis, and increased inflammatory markers (CRP).
Remember this for later.
(Image)
Participants received flucloxacillin 1000 mg IV every 6 hours, and then therapy was stepped down to flucloxacillin 500 mg PO every 6 hours. Total duration of therapy was either 6 or 12 days of therapy, randomized in a 1:1 ratio on day 5 of therapy. In order to be randomized, patients had to have improved after 5 days, remained afebrile, been on the same antibiotic, maintained negative blood cultures, and not had a change in diagnosis.
Follow-up was performed at 14, 28, and 90 days. Adherence was assessed with a pill count at day 14 or by telephone calls if patients were unable to attend. Follow-up was also done by telephone on day 90.
The primary outcome was cure by day 14 without relapse by day 28. Cure by day 14 was defined as absence of warmth and pain with reduction in erythema. Relapse was defined as initiation of antibiotics for cellulitis by non-trial physicians.
Thirty-eight patients (50%) were considered cured without relapse in the 12-day group. This outcome occurred in 36 patients (49.3%) in the 6-day group (95% CI -15.0 to 16.3; P = 1) meaning there was no significant difference between groups.
(Image)
But check out that confidence interval. It is WIDE, which means there was high variability among patients in both groups.
Another important point is the fact that this study was underpowered. The authors needed 178 patients per group to have a 10% non-inferiority margin. Between both groups, the investigators were able to recruit only 149 patients in total.
As you may recall, studies are powered to reduce the probability of having a type II error (aka to avoid a false negative - meaning studies should be powered to avoid missing a difference when there actually is one).
Based on the fact that this study was underpowered and the confidence interval had a wide margin, the results of this study should be taken with caution. Maybe there actually was a difference between the 2 treatment durations and the study just didn’t have enough enrollees to detect it!
Plus, let’s have a moment for those cure rates, regardless of group. Cure rates reported in the DANCE trial were near the 50% mark, whereas similar studies have reported cure rates higher than 90%.
Would you want to take a medication that only worked 50% of the time? Not so hot, right?
Remember that the inclusion criteria weren’t overly robust in the objective infection arena, which may have meant patients with non-infectious processes, such as dermatitis and venous stasis, were also included in the study. This could have contributed to the low cure rates reported in this study - because maybe the problem wasn’t an infection to begin with!
Have I told you guys about how important secondary outcomes are when interpreting a clinical trial? I think I did in part one…but just as a refresher, secondary outcomes provide lots of information when interpreting studies.
Recurrence rates (defined here as prescription of antibiotics by non-trial clinicians) at 90 days were higher in the 6-day treatment group vs. the 12-day treatment group (23.5% vs 5.7%, 95% CI -34.8 to -0.8, P-value = 0.045). Once again, when interpreting this we should take into consideration the wide confidence interval, the fact this was a subjective outcome measure (just because you get antibiotics doesn’t prove you have infectious cellulitis), and the low follow up response rates (only about half of patients responded to phone calls).
As a matter of fact, the confidence interval margin was so wide that if only one more patient in the 12-day group reported a relapse, the difference for relapse at day 90 would have been non-significant.
Man…#STATS. Convenient stats. Just sayin’.
Combined with the fact that only half of the patients responded and that this measurement was subject to what the patient told the investigators over the phone (remember recall bias), this secondary outcome is on seriously shaky ground.
When taking a closer look at the baseline characteristics, we can see some differences between groups, which is even more important in a trial with a small number of participants - like this one. Some of the most notable baseline differences between groups are patient weight and previous cellulitis episodes.
Patients in the 6-day group had a BMI of 33 kg/m2 vs a BMI of 29 kg/m2 in the 12-day group. Also, more patients in the 6-day group had at least one previous episode of cellulitis when compared to the 12-day group, which meant increased previous antibiotic exposure in the 6-day group.
Obesity and previous episodes of cellulitis may be associated with an increased risk of recurrent cellulitis. Additionally, previous exposure to antibiotics may be associated with an increased risk of developing MRSA infections, which is important to consider when talking about selection of antimicrobials.
This brings us to antibiotic choice in the study.
The choice of antimicrobial for this trial was flucloxacillin. Flucloxacillin is a penicillinase-resistant penicillin (aka an anti-staphylococcal penicillin) which has activity against Staphylococcus and Streptococcus species. Flucloxacillin does not have activity against MRSA.
It is recommended for cellulitis in the European CREST guidelines for the management of cellulitis, but these guidelines do differ from the IDSA guidelines on skin and soft tissue infections. In contrast to the IDSA guidelines, the CREST guidelines do not recommend adding MRSA coverage, even in patients with SIRS or who have failed oral antibiotics.
Let’s look at that one-third of patients in the study who received previous antibiotics for the treatment of cellulitis. According to IDSA guidelines, these patients would have been classified as having severe cellulitis given their failure of PO antibiotics, and the recommendation would have been for MRSA coverage. But in the study, patients received flucloxacillin - aka no MRSA coverage.
So perhaps this lack of MRSA coverage may have contributed to the poor response rates reported in this study! The choice of antibiotic limits the external validity (or generalizability) of this study’s results for our patients here in the US.
To wrap up the DANCE trial…
Currently, the IDSA guidelines recommend a treatment duration of 5 days in non-purulent cellulitis. Despite many clinicians treating longer, this study likely isn’t going to be the turning point for shortening cellulitis treatment practices. There were SO many limitations, including:
Limited follow up
Important differences in baseline characteristics
Lack of guideline and medication applicability
Controversy in the definition used for cellulitis
Questionable methods to assess recurrence at 90-days, which may have had some recall bias and poor participation.
All of these factors limit the utility of this study in supporting IDSA’s cellulitis treatment recommendation for a 5-day course of antibiotics guided by patient response.
So there you have it! ID updates in a nutshell. Hopefully this information will spur you to read these studies yourselves and see what you think, but at the very least, you have the big takeaway points here. I look forward to updating you soon with some more clinical trial info!!