Overview of mold and IMIs

Mold is widely prevalent indoors and outdoors and can travel through the air and persist on surfaces. In particular, molds grow well on materials that remain moist for at least 24–48 hours.^{Reference Lockhart, Pham and Gade14} Molds do not cause health problems for most people, but for some, they can cause health conditions ranging from allergies or asthma to IMIs. IMIs are serious infections that can cause substantial morbidity and mortality.^{Reference Kanamori, Rutala, Sickbert-Bennett and Weber1,Reference Enoch, Yang, Aliyu and Micallef7,Reference Vallabhaneni, Benedict, Derado and Mody15,16} Of the hundreds of thousands of species of molds, only a small number cause most IMIs, primarily Aspergillus species (particularly A. fumigatus) and mucormycetes, most of the genera Rhizopus, Mucor, and Rhizomucor. IMIs are associated with an estimated 16,000 hospitalizations and $1.4 billion in direct medical costs in the United States each year, as well as mortality rates as high as 50% in persons with immunocompromising conditions.^{Reference Vallabhaneni, Benedict, Derado and Mody15,Reference Benedict, Jackson, Chiller and Beer17-Reference Maschmeyer, Calandra, Singh, Wiley and Perfect19} Mucormycosis (infections caused by mucormycetes) most commonly occur in immunocompromised patients. Invasive aspergillosis (infections caused by Aspergillus) also occurs in immunocompromised patients, but cases have also been reported in immunocompetent patients in the intensive care unit (ICU) or following surgery.^{Reference Taccone, Van den Abeele and Bulpa20-Reference Patterson, Thompson and Denning24} One study reported that ∼10% of invasive aspergillosis cases occurred in postsurgical patients.^{Reference Jensen, Guinea, Torres-Narbona, Muñoz, Peláez and Bouza25}

Overview of IMI clusters and outbreaks

The annual number of IMI clusters and outbreaks in the United States is unknown, primarily because IMIs are usually not reportable to public health in this country through systematic surveillance. In the United States, only Seattle–King County, Washington, has made IMI cases reportable (starting in February 2020), and aspergillosis is reportable in Louisiana. Therefore, routine baseline data about IMI frequency are often unavailable. Given the absence of public health surveillance for IMIs, most IMI investigations arise when a healthcare facility identifies a cluster of cases that share a common location, population, or time period.

An IMI cluster is a group of 2 or more patients with IMI that may signal an ongoing outbreak but has either not been fully investigated or lacks sufficient data to confirm the existence of an outbreak. Because of the challenges inherent in confirming IMI outbreaks, clusters are important signals that often prompt further investigation. An IMI cluster is considered an IMI outbreak when the number of infections is confirmed to be larger than expected within a defined population (ie, certain high-risk patients) over a specific period (e.g., 6 months), or when 2 or more cases are linked to a common source. An increased incidence over an estimated baseline rate or the presence of IMI in patients who are not considered at highest risk should also prompt evaluation by public health and hospital personnel.^{Reference Patterson, Thompson and Denning24} In addition, a case of IMI with suspected hospital exposure (e.g., admission >7 or 14 days before presumed IMI onset or one that is associated with surgery) should prompt evaluation. Patient notification should be considered in both clusters and outbreaks because clusters may later be confirmed as outbreaks and early notification is key to helping reducing risk to patient health; here, we refer to them collectively as clusters for simplicity.

Because IMI clusters often involve a small number of patients and mold exposure can occur in many settings, identifying the exposure source is often difficult. Confirmed sources have included construction or demolition sites, water damage, and ventilation system deficiencies.^{Reference Kanamori, Rutala, Sickbert-Bennett and Weber1,Reference Chow, Toda and Pennington18,Reference Rammaert, Lanternier and Zahar26-Reference Pini, Faggi, Donato, Sacco and Fanci30} Less common sources include hospital linens, medications, foods, and medical supplies and equipment.^{Reference Hartnett, Jackson and Perkins3,Reference Duffy, Harris and Gade31-Reference Lalayanni, Baliakas and Xochelli35} Even if the source cannot be identified, all investigations offer opportunities for healthcare facilities to inspect and remediate possible sources to improve patient safety. When a US healthcare facility identifies an IMI cluster, it should notify its local or state public health department; health departments can contact CDC (fungaloutbreaks@cdc.gov) for additional assistance. Additionally, an industrial hygienist can also be contracted to assist with the investigation in the hospital setting.

Healthcare-associated IMI clusters can attract substantial media attention. Recent examples include a 2014–2015 mucormycosis outbreak in Pennsylvania and a 2018–2019 aspergillosis outbreak in Washington state.^36,37 Media coverage of these events represent one of the few opportunities for the general public to learn that molds can cause infections.

Communicating with the public and the media about IMI clusters can be challenging, particularly because these clusters involve severe infections, can be difficult to confirm, and often involve unclear sources. Communication strategies employed in other contexts, such as breaches in infection control and injection safety,^{9,Reference Schaefer, Perkins, Link-Gelles, Kallen, Patel and Perz13} can be applied to IMI clusters, though additional considerations are often appropriate. One such factor is that mold may be perceived as associated with visibly unhygienic conditions, even though mold exposures can occur in the absence of visible mold. Additionally, IMI clusters can be particularly challenging to investigate, complicating patient notifications. We briefly summarize these challenges below.

Challenges to identifying healthcare-associated IMI clusters

Several features of IMIs make them particularly challenging to investigate, not least of which is that many infection control staff lack experience dealing with IMI clusters. IMI signs and symptoms are often nonspecific and sometimes insidious at the outset, making it difficult to establish an onset date in many cases. In addition, the incubation period for several IMIs, particularly noncutaneous mucormycosis, is not well defined,^{Reference Hartnett, Jackson and Perkins3,Reference Messina, Wolfe and Hemmersbach-Miller38} largely because it is difficult to ascertain an exposure date. For this reason, investigators typically use a wide exposure window (e.g., 14 days or longer) when evaluating exposures as potential sources. Diagnosis of an IMI typically requires invasive specimens for culture or histopathology; adjunctive tests like galactomannan, β-D-glucan, and PCR testing can provide supportive evidence. Given the multiple testing modalities, creating case definitions and performing laboratory-based case finding can be complex. No simple test exists for reliably diagnosing IMI^{Reference Kelly, Pennington and Limper39,Reference Farges, Cointault and Murris40} because a positive laboratory test can signify infection or colonization, necessitating additional contextual information in many cases. IMIs are among the most common missed diagnoses among ICU patients;^{Reference Winters, Custer and Galvagno41} delays in diagnosis can increase the severity and mortality of IMIs. Even thorough IMI investigations may fail to identify a source, particularly for small clusters, which have few cases to provide epidemiologic clues.^{Reference Hartnett, Jackson and Perkins3} The list of potential mold sources is lengthy, as described above, and mold can be difficult to detect in the environment unless visible contamination is present.

Challenges to environmental sampling during suspected healthcare-associated IMI clusters

Environmental sampling is often seen as the critical step in identifying a source of an IMI outbreak, but its utility is often overestimated. It is best performed using a well-designed sampling plan in combination with a thorough environmental assessment, sometimes under the guidance of an industrial hygienist, to most effectively identify the mold source. Other important considerations include environmental sampling equipment, experience of the person performing the air sampling, and the expertise of the microbiology lab. Conventional microbiologic methods may not always accurately speciate molds, which is important when comparing environmental sampling results with those from patients. Additionally, the diagnostic criteria needed to distinguish invasive infection from colonization are complex, and mycology laboratory capacity for species determination can vary by hospital, often leading to misidentifications in clinical laboratories.^{Reference Hartnett, Jackson and Perkins3}

In addition to laboratory challenges, environmental testing results can be difficult to interpret. An environmental sample positive for mold can help identify potential targets for remediation but such a finding cannot be interpreted as indicating a definitive source of transmission in the absence of an epidemiologic link to patients.^{Reference Hartnett, Jackson and Perkins3} A negative result is limited in that it indicates that spores were not present at the specific time of sampling using a specific method, but such results cannot be used to rule out the presence of mold because mold may not be consistently present.^{Reference Hartnett, Jackson and Perkins3} Furthermore, because environmental conditions can change between time of exposure and when sampling is conducted, a negative test does not exclude previous contamination of the sampled site.

Considerations for patient notification about IMI clusters and outbreaks

Methods for and examples of using patient notification in IMI outbreaks

Notification methods can include mailed documents, electronic messages, in-person discussions, and telephone calls. Telephone calls or in-person discussion are preferred if the number of patients being notified is relatively small, and telephone calls may be most feasible for patients who have already been discharged from the facility.^{9,12,Reference Schaefer, Perkins, Link-Gelles, Kallen, Patel and Perz13} Following telephone calls and in-person discussions, written information should be provided to patients so they have easily accessible and accurate documentation to refer back to or share with caregivers and other healthcare providers.

Recently, general guidance for emergency and risk communication strategies for suspected outbreaks have become available and include the CDC Introduction to Patient Notification Toolkit⁹ and Crisis and Emergency Risk Communication Manual.¹⁰ The Council for Outbreak Response: Healthcare-Associated Infections and Antimicrobial-Resistant Pathogens’ (CORHA) Interim Framework for Healthcare-Associated Infection Outbreak Notification also provides guidance for notification in the context of a suspected healthcare-associated infection (HAI) outbreak.¹²

The following scenarios describe fictional examples of patient notification during IMI cluster investigations intended to represent how notification can affect outcomes. They do not reflect actual events.

1. Hypothetical aspergillosis cluster. A hospital identified 6 cases of mucormycosis (3 Rhizopus and 3 Mucor spp.) among ICU patients receiving mechanical ventilation over the previous 4 months, none of whom were immunosuppressed, apart from in-hospital corticosteroid use. Five had been hospitalized for >2 weeks before suspected mucormycosis onset. For comparison, only 2 cases of mucormycosis had been diagnosed in ICU patients over the previous 3 years. A team at the hospital began to investigate by reviewing patient records and taking air and surface samples for mold spore identification. The air and surface samples identified a relatively low level of mold spores (although no specific guidelines exist for healthcare settings), with few spores suggestive of mucormycetes. The team suspected that the cluster of cases might have reflected increased use of diagnostics to detect mucormycosis; no patient notification was performed.

Over the next 3 months, 5 more cases of mucormycosis were identified in ventilated patients. Environmental sampling was performed again, with similar results. One patient’s family contacted a local television news station, which aired a story about deadly mold infections at the hospital, prompting concern from many patients and families. National news coverage followed.

In consultation with the state health department and an industrial hygienist, the hospital performed a thorough environmental assessment and identified a likely source: water-damaged building materials in a wall of the ventilator storage area. Previously unidentified gaps in the room wall, along with unbalanced air flow, allowed dust into the clean storage area. Targeted culture-based air and surface samples identified high levels of mucormycetes, and the area was successfully remediated.

2. Hypothetical mucormycosis cluster. A hospital notified their local health department after identifying three cases of aspergillosis over the previous 3 months in inpatients with hematologic malignancies. Two of these patients had been inpatients for >14 days before the suspected aspergillosis onset date, suggestive of healthcare-acquired infections. On a retrospective review of microbiology and pathology data, 0–1 cases per year had been identified over the past 5 years. Based on the cluster epidemiology and knowledge of high-risk conditions for IMI, healthcare personnel created a list of patients who had been treated for hematologic malignancy at this hospital in the month before the first cluster case. Patients, or family members when appropriate, on this list were notified at in person-visits or by telephone and were provided a letter and factsheet that described the cluster, the patients’ potential health risks, general information about IMIs, and recommendations for patients and visitors entering the facility. Despite thorough investigation, the source was not identified, but based on an environmental assessment, the hospital made several improvements to the ventilation system, cleaning procedures, and linen storage. The infection control team also instituted new policies for routine assessment of potential mold sources, including meetings with building engineers and maintenance.

Communicating early and proactively about potential healthcare-associated IMI outbreaks is key to helping reduce additional risks to patients’ health. Early notification can allow patients to feel more informed about their health status and helps foster trust between patients and healthcare facilities. Successful patient notification relies on close collaboration between healthcare facility personnel, public health, and communication professionals. IMI outbreaks often involve many scientific uncertainties and challenges, especially in identifying the source. More research is needed on how and why healthcare-associated mold outbreaks occur and on IMIs in general. Insight into the general public’s knowledge and perceptions about mold as it relates to health is also needed. Establishing regional or nationwide surveillance for IMIs would also help establish the true prevalence of these infections and help identify baseline rates and outbreaks. In summary, despite the challenges associated with identifying and communicating about healthcare-associated mold outbreaks, patient notifications are an important part of the response to IMI outbreaks, allowing patients and providers to make informed care decisions.