Which Multidrug-Resistant Organisms (MDROs) Cause High Death Rates?

Carbapenem-resistant gram-negative bacteria consistently top the list of deadliest multidrug-resistant organisms, with mortality rates in intensive care settings regularly exceeding 40% and sometimes climbing past 70%. The organisms that kill at the highest rates are not necessarily the ones the public hears about most. MRSA gets the headlines, but carbapenem-resistant Acinetobacter baumannii, carbapenem-resistant Enterobacterales, and drug-resistant Candida auris are, in many clinical settings, substantially more lethal. The picture is complicated by the fact that the same bug can be a minor nuisance in one patient and a death sentence in another, depending on infection site, immune status, and how quickly the right treatment starts.

Carbapenem-Resistant Acinetobacter baumannii

If there is a single organism that infectious disease specialists dread, it is carbapenem-resistant Acinetobacter baumannii, or CRAB. This bacterium thrives on hospital surfaces and ventilator circuits, and once it becomes resistant to carbapenems, the strongest broad-spectrum antibiotics available, treatment options shrink dramatically. A prospective study of fifty ICU patients with confirmed CRAB infections found that 40% died, and clinical improvement occurred in only half. The study authors described frequent panresistance, meaning the bacteria were resistant to essentially every drug tested.¹

That 40% figure may actually understate the problem. A study in a Moroccan teaching hospital reported crude mortality of about 74% among ICU patients with A. baumannii infections, with septic shock and older age being the strongest predictors of death.² The range reflects real differences in patient populations and available therapies, but even the lower end is sobering. CRAB is particularly devastating because it often strikes patients who are already critically ill, on ventilators, or recovering from major surgery, and the treatment toolkit for it remains thin.

Carbapenem-Resistant Enterobacterales

Carbapenem-resistant Enterobacterales, or CRE, is a family of gut bacteria that includes Klebsiella pneumoniae, Escherichia coli, and Enterobacter species. The CDC once dubbed CRE “nightmare bacteria,” and the mortality data support the label. A multicenter study of patients with CRE bloodstream infections found an overall 30-day mortality rate of about 51%.³ Other studies place 30-day mortality in the range of 39% to 42%, with sepsis at the time of diagnosis, pneumonia as the infection source, and inappropriate initial antibiotic choice all sharply increasing the odds of death.⁴⁵

Klebsiella pneumoniae deserves special mention within the CRE group. Some strains carry “hypervirulence” genes that help the bacteria evade immune defenses and invade tissues more aggressively. When hypervirulence and carbapenem resistance combine in the same strain, the median mortality sits around 30%, though the range across published studies is wide.⁶ Interestingly, head-to-head comparisons of hypervirulent versus classical carbapenem-resistant Klebsiella have not always found a statistically significant difference in death rates, suggesting that carbapenem resistance itself is the dominant driver of mortality rather than the extra virulence genes alone.⁷

Drug-Resistant Pseudomonas and Stenotrophomonas

Pseudomonas aeruginosa is an opportunistic pathogen with an inherent talent for resisting antibiotics. It causes ventilator-associated pneumonia, wound infections, and bloodstream infections, particularly in burn units and ICUs. The carbapenem-resistant phenotype roughly quadruples the odds of dying compared to susceptible strains, and both carbapenem-resistant and multidrug-resistant phenotypes are linked to higher early mortality within the first five days of bloodstream infection.⁸ Looking at longer-term outcomes, a cohort study of patients with multidrug-resistant P. aeruginosa in the urinary tract found an overall mortality of about 17%, with roughly a third of those deaths estimated to be directly tied to the resistant infection.⁹

Stenotrophomonas maltophilia is less well known but increasingly recognized as a serious threat, especially in patients with blood cancers or those on prolonged ICU stays. This organism is intrinsically resistant to carbapenems and many other drug classes, making it difficult to treat from the outset. Studies report 28-day mortality rates for S. maltophilia bloodstream infections ranging from about 35% to 46%.¹⁰¹¹ In patients with blood disorders, prolonged neutropenia and ceftazidime resistance independently worsen the prognosis.¹²

MRSA and Its Lingering Toll

Methicillin-resistant Staphylococcus aureus is probably the most publicly recognized MDRO, and it remains a major killer, though its headline mortality rates are lower than those of the gram-negative organisms described above. Within a year of diagnosis, roughly 22% of patients with community-diagnosed MRSA infections died, compared with 5% of patients with non-MRSA Staphylococcus aureus, giving MRSA patients about four times the risk of death after adjustment.¹³

What makes MRSA distinctive is how much of its mortality is delayed. A study tracking patients after their initial Staph infection found that at three months post-discharge, MRSA patients had a mortality rate of 32% versus 18% for those with methicillin-susceptible strains. By twelve months, the gap had widened further: 51% versus 32%.¹⁴ Most deaths occurred after hospital discharge, not during the initial stay. This long tail of excess mortality is easy to miss when studies only track in-hospital outcomes, and it means that MRSA’s true burden is larger than the acute mortality numbers suggest.

Vancomycin-Resistant Enterococci

Enterococci are gut bacteria that occasionally cause bloodstream infections, urinary tract infections, and infections of the heart valves. When they develop vancomycin resistance, especially the species E. faecium, mortality climbs substantially. A systematic review found that 30-day mortality for vancomycin-resistant enterococcal bloodstream infections ranged from about 28% at the low end to 70% at the high end, depending on the patient population and study setting.¹⁵ Compared head to head with vancomycin-susceptible strains, resistance nearly doubles the hazard of dying.¹⁶

VRE disproportionately affects patients in oncology wards, transplant units, and long-stay ICU beds. One study found 28-day mortality of about 51% among patients with vancomycin-resistant E. faecium bacteremia.¹⁷ The bacteria themselves are sometimes less virulent than gram-negative pathogens, but the patients they infect tend to be profoundly immunocompromised, and the loss of vancomycin as a treatment option leaves clinicians relying on drugs with narrower effectiveness and more side effects.

Candida auris

Not all deadly MDROs are bacteria. Candida auris, a yeast first identified in Japan in 2009, has since spread to hospitals on every inhabited continent and quickly earned a reputation as one of the most dangerous hospital-acquired fungal infections. It resists multiple classes of antifungal drugs, persists on surfaces for weeks, and spreads readily between patients. In a large U.S. hospital database covering 2017 to 2022, the estimated crude mortality for C. auris-associated hospitalizations was 34%.¹⁸

Published estimates of mortality from C. auris bloodstream infections range widely, from about 29% to 62%, depending on the setting.¹⁹ One multicenter cohort that compared C. auris directly with other Candida species found that after adjusting for treatment and septic shock, C. auris bloodstream infection actually carried somewhat lower mortality than infections from other Candida species combined.²⁰ The raw numbers look high partly because C. auris tends to infect the sickest patients in the hospital. That does not make it less dangerous in practice, since those are exactly the patients who can least afford an untreatable infection, but it does mean that the organism itself may not be inherently more lethal than other Candida species once you account for how fragile its hosts already are.

Multidrug-Resistant Tuberculosis

MDR-TB belongs in a separate conversation from the hospital-acquired organisms above, because tuberculosis is a community-transmitted disease with a treatment course measured in months or years rather than days. But its death toll is enormous. Globally, antimicrobial resistance across all pathogens was responsible for roughly 1.14 million deaths directly attributable to resistance in 2021, with millions more associated deaths.²¹ MDR-TB is one of the single largest contributors to that total.

Compared with drug-susceptible tuberculosis, MDR-TB multiplies the risk of dying during treatment roughly sevenfold.²² A long-term follow-up study in Wuhan, China found that about 40% of MDR-TB patients died during follow-up, with older age, lower education level, and treatment interruption all raising the risk substantially.²³ In settings with access to newer short-course regimens, outcomes can be better: an Indonesian study reported that roughly 9% of MDR-TB patients on short-course treatment died, though older patients still faced markedly higher risk.²⁴ The gap between 9% and 40% illustrates how much treatment access and adherence shape outcomes in MDR-TB, far more than for many hospital-acquired MDROs where even optimal therapy may not save the patient.

Why Resistance Makes Infections So Much Deadlier

The obvious explanation is that when the usual drugs do not work, doctors have fewer options and some patients get no effective treatment at all. But the mechanism goes beyond just running out of antibiotics. A big part of the mortality penalty comes from timing. In sepsis, every hour of delay in receiving an effective antibiotic increases the odds of dying. A meta-analysis found that each hour of delayed antibiotic treatment raised in-hospital mortality by about 4%, and giving antibiotics after the first hour rather than within it raised mortality by about 20%.²⁵ In patients with septic shock, a large multicenter study estimated that each hour of delay added about 1.8 percentage points to absolute mortality.²⁶

When a patient has an MDRO infection, the empiric antibiotics given at admission, chosen before lab results come back, are more likely to be the wrong ones. By the time cultures confirm which organism is present and which drugs still work, hours or days have passed. This gap between infection onset and effective treatment is the single biggest mechanism by which resistance kills. In multiple CRE studies, inappropriate empiric therapy was independently associated with dying, with one analysis showing it roughly quintupled the odds of death.⁴

The remaining drugs for highly resistant infections also tend to be more toxic, harder to dose correctly, and less effective. Colistin, one of the last-resort antibiotics for gram-negative MDROs, damages the kidneys. Linezolid, used against VRE and MRSA, suppresses bone marrow. Patients are caught between the infection and the side effects of salvage therapy.

Populations at Greatest Risk

MDROs hit hardest in people whose immune defenses are already weakened. Solid organ transplant recipients are a stark example. A meta-analysis found that colonization with multidrug-resistant bacteria before or after transplant more than doubled the risk of dying within a year. Colonization with CRE was the most dangerous, nearly quadrupling one-year mortality, while MRSA colonization roughly doubled it.²⁷ Liver transplant recipients seem particularly vulnerable, partly because the surgical site is close to the gut where resistant bacteria reside.²⁸

Newborns in intensive care are another highly susceptible group. A study comparing neonates with multidrug-resistant gram-negative sepsis to those with susceptible strains found that case fatality was roughly 3.6 times higher in the resistant group, and septic shock rates were dramatically elevated.²⁹ Premature infants, whose immune systems are underdeveloped and who often need prolonged IV access, are at especially high risk for acquiring these organisms.

COVID-19 hospitalization revealed another dimension of vulnerability. In a study of nearly 500 patients with COVID-19 pneumonia, those who developed secondary hospital-acquired infections from MDROs had in-hospital mortality of about 84%, compared with roughly 22% for COVID-19 patients without MDRO superinfections.³⁰ This fourfold difference underscores how MDRO infections pile on top of existing critical illness, creating a compounded threat that overwhelms even aggressive supportive care.

New Treatments and Their Limits

The antibiotic pipeline has not been entirely dry. Newer beta-lactam/beta-lactamase inhibitor combinations like ceftazidime-avibactam have become important weapons against CRE, especially Klebsiella pneumoniae. A single-center study of 142 patients with carbapenem-resistant K. pneumoniae infections treated with ceftazidime-avibactam found a 28-day survival rate of 78%.³¹ That represents a meaningful improvement over historical mortality rates for CRE, though more than one in five patients still died.

The problem is that resistance follows treatment. In one study of carbapenem-resistant Klebsiella pneumoniae from ventilator-associated pneumonia, 79% of isolates were already resistant to ceftazidime-avibactam. A newer drug, cefiderocol, maintained activity against 96% of the same isolates, offering a backup option, but its long-term resistance trajectory remains uncertain.³² The pattern is familiar: a new drug arrives, works well for a few years, and then resistance begins to erode its effectiveness. The pharmaceutical development cycle for antibiotics is slow relative to how quickly bacteria adapt.

What Happens When Every Drug Fails

Pan-drug-resistant organisms, those resistant to every available antibiotic, are no longer hypothetical. They appear sporadically in ICUs worldwide, and modeling studies have tried to estimate the consequences of their broader spread. A modeling study focused on E. coli projected that mortality from a hypothetical pan-resistant strain causing sepsis could reach anywhere from 34% to 96% in adults over 65, depending on assumptions about supportive care and host factors.³³ Even the low end of that range is devastating for what is currently a routinely treatable bacterium.

In clinical practice, a retrospective ICU study comparing patients with extensively drug-resistant or pan-drug-resistant gram-negative infections to those with less resistant strains found longer ICU stays, more vasopressor use, and higher raw mortality. Yet resistance phenotype alone did not independently predict death in the multivariate analysis, suggesting that the severity of the underlying illness still matters more than the resistance profile in determining who survives.³⁴ That finding is a reminder that resistance and mortality are tangled up with how sick the patient already is. The most resistant organisms tend to appear in the most debilitated patients, making it hard to separate how much of the death rate is driven by the bug versus the host.

Long-Term Care and the Mortality That Gets Missed

Much of the research on MDRO mortality focuses on acute hospitalization, but a substantial portion of the burden unfolds later, in rehabilitation facilities, long-term care, and after discharge. MRSA patients, as noted earlier, continue dying at elevated rates for a full year after their initial infection. A prospective study in a long-term care facility found that mortality among MDRO-colonized residents depended more on sputum carriage and the resident’s overall health than on which specific resistant organism they carried.³⁵ In other words, simply carrying a resistant organism in the gut or on the skin is not itself a death sentence, but once it causes active infection in a frail host, the outcomes are poor and the ripple effects extend well beyond the initial hospital stay.

This post-acute mortality is systematically undercounted. Hospital databases capture in-patient deaths, but a patient discharged alive after CRE bacteremia who dies three weeks later at a skilled nursing facility may never appear in the hospital’s mortality statistics. Studies that follow patients for 90 days or a year consistently find higher overall death tolls than 30-day studies, which means the commonly cited 30-day mortality rates for most MDROs, while already alarming, underestimate the full impact of these infections on survival.