Below you will find the recap of the 1st case in our Sepsis Smackdown case series. Several weeks ago, we posed a clinical scenario followed by a series of questions. As a refresher the case was that of Linda, a 79 yo female resident of a local nursing facility who arrives to your busy community ED with altered mental status, hypothermia, tachycardia, hypotension, and along history of multiple complicated UTI's. You begin your work up and find her to again have a likely UTI with a urine dipstick with large leukocyte esterase, pH 5.5, 20 protein, negative nitrite.
Q1: What antibiotic(s) would you choose at this point? Does a remote history of one serious UTI change what you would use?
In responding to this question, several providers wanted to point out the potential cognitive pitfalls of anchoring and premature closure with this patient. It is still uncertain if this patient’s presentation is clearly due to a urinary tract infection. With an unclear diagnosis, most felt it appropriate to start broad spectrum coverage (typically meropenem), irrespective of what the UA showed, or what the more likely source might be. While there is often a lot of discussion about time to antibiotic coverage in sepsis, less emphasis than is warranted is placed on appropriate coverage. When we limit our differentials and antibiotic coverage in the ED, it carries a lot of momentum into the patient’s ICU/inpatient stay. If we inappropriately limit our antibiotic choices, it could cause bad outcomes for the patient down the road. The inpatient team can always narrow down coverage later, based on culture results. Even with the potential adverse outcomes of providing too much coverage (allergic reactions, diarrhea, kidney damage, etc), it is better to cover the bug now, then to be behind the 8 ball three days later when the cultures turn positive for something else.
Presuming that the rest of the workup did not find another more likely source of sepsis, most felt that going bigger was better on this patient, for a variety of reasons. Meropenem was preferred, due to her previous pseudomonal UTI resistant to other drugs, in combination with her severe state of sepsis.
This question also served to prompt a discussion on interpretation of urinalyses. This patient was nitrite negative. This test usually detects anaerobic gram negative bacteria that are able to reduce nitrate into nitrite. While generally very specific for UTI, the nitrite test is only about 35-85% sensitive. Sensitivity can be bolstered, however, by evaluation of leukocyte esterase, which is much more sensitive for presence of a UTI. Absence of nitrites does not indicate a urine not infected by anaerobic gram negatives, however, as there is often a threshold colony count required to yield a positive on a urine dipstick, this may not always be met in all UTIs.
Q2: Now, imagine the same patient, but with a perfectly clean urinalysis, clear chest x-ray, and a benign belly. Lactate 6.5 and creatinine 2.3 (baseline 1.6), but nothing else particularly remarkable. Oh, and she’s HIV+, but last CD4 count 6 months ago was 405. What would your choice of antibiotics then be? Would your coverage change if she were not so well treated for her HIV?
There was little specific response to this question. The focus here was placed on avoiding inappropriate treatment momentum, and keeping a broad differential, especially in HIV patients. We should consider options outside of infectious causes in these patients with shock and hypotension (status epilepticus, thyroid disorders, vascular emergencies, etc). In addition to keeping a broad differential, the focus here should perhaps have been on providing additional coverage and doing a workup for opportunistic infections in uncontrolled HIV. In patients with HIV and a favorable CD4 count, it is unlikely that additional coverage will be needed beyond the standard “broad spectrum” of choice. While we don’t always have a CD4 count available in the ED, it can be estimated by calculating the absolute lymphocyte count (ALC) from the CBC and differential (WBC x percentage of lymphocytes). Evidence shows a high PPV for a CD4 count less than 200 when the ALC is less than 1000, and a high NPV for an ALC greater than 2000. Counts between 1000 and 2000 are somewhat ambiguous (Absolute lymphocyte count as a predictor of CD4 count. Ann Emerg Med. 1998 Sep; 32:323-8.)
Q3: Now instead of HIV, the patient on tacrolimus and prednisone for a previous heart transplant. When calling her up to the ICU residents, you ask them what they think about adding on antifungals or antivirals. The response is, “nah, don’t worry about it.” What is your threshold for adding additional coverage? Would your opinion on the matter change if it were her transplant physician that told you not to worry about adding antifungals or antivirals?
There was a fair amount of concern for a fungal infection in this type of patient. Most responders felt somewhat inclined to initiate treatment with an antifungal based upon risk factors alone. However, it is noted that most fungal infections tend to be more indolent and long-term (requiring long-term treatment), and that if an inpatient team did not deem it worthy to continue an antifungal treatment that was initiated in the ED, there would be little utility to initiating it at all. Most felt that if the plan wouldn’t be carried forward as an inpatient, then there was no use in providing antifungal coverage. There did not seem to be much of a difference in this opinion depending on whether the recommendation came from an attending provider or a resident. However, some who replied to this question were of the opinion that we should be held to the same standard of treatment with antifungals and antivirals as we are to antibiotics in these high risk patients, and that we should treat them urgently and provide a dose in the ED regardless of other opinions. In either event, if not initiating treatment with antifungals or antivirals, it is appropriate to send fungal and viral cultures.
Fungal infections are more common in patients with organ transplants. Candida and aspergillus are the most common offending organisms, and guidelines recommend aggressive care for transplant patients diagnosed with an invasive fungal infection. Risk factors for invasive aspergillus infections include AML, stem cell transplants, solid organ transplants, and those on prolonged immunosuppression.
Q4: What if this patient with an uncertain source of sepsis had had a kidney transplant and visible thrush? How do you weigh the consequences of potentially causing renal injury by adding IV fluconazole (or acyclovir)?
Similar to what was discussed above, concerns about potential fungal infections in transplanted patients were echoed. In renal transplant patients invasive fungal infections represent between 1.4% and 9.4% of all infections, nearly half of these caused by candida. Beyond that, aspergillus and cryptococcus are the most common offending organisms. Most appropriately felt that in a transplant patient with signs of candida infection and sepsis, that intravenous antifungals should be initiated. While there are concerns for kidney (micofungin, acyclovir) or liver (micofungin, acyclovir, and fluconazole) injury or failure with treatment of these atypical infections, the current suspicion for fungal infection in this patient outweighs the potential risk of organ damage from treatment.
Author: Nathanial Mann, MD, PGY-4 UC EM Residency
Edited: Jeffery Hill, MD MEd