Invasive mold infections following Hurricane Harvey—Houston, Texas

Publication summary

Invasive mold infections are responsible for about 16,000 hospitalizations and $1.4 billion in medical costs in the US every year. Extreme weather events precipitated by climate change may increase the risk of environmental mold inhalation and affect populations who have not historically been at risk of invasive fungal infections. During August and September 2017, Hurricane Harvey dropped more than 40 inches of rain on Houston, Texas, USA, resulting in extreme flooding and consequent mold exposure for residents and clean-up workers. This study assesses the incidence of invasive mold infection at four hospitals—one of which is a cancer care facility—before and after Hurricane Harvey made landfall in Houston, Texas.

Who this is for

• Epidemiologists, particularly in areas that experience extreme weather and flooding events
• Infectious diseases clinicians and medical mycologists
• Environmental policymakers

Key findings

Invasive mold infection incidence

The researchers identified 182 patients with invasive mold infections treated at four Houston, Texas, hospitals from Sep 1, 2016, to Aug 31, 2018 (Pre-Hurricane Harvey: Sep 1, 2016–Aug 31, 2017; Post-Hurricane Harvey: Sep 1, 2017–Aug 31, 2018). 

Among patients with invasive mold infections, 55 cases were proven (i.e., positive culture from a normally sterile site or evidence of tissue invasion on histopathology), 41 cases were probable (i.e., one patient-associated risk factor for invasive mold infection, a clinical feature of infection such as radiologic abnormalities, and evidence of fungal presence such as a sputum culture), and 86 cases were surveillance (i.e., receipt of mold-active systemic or ocular therapy and one or more patient-associated risk factor for invasive mold infection).

The incidence of invasive mold infection increased after the hurricane, reaching 3.69 cases per 10,000 healthcare encounters (109 cases/295,443 encounters), compared with 2.5 cases per 10,000 encounters (73 cases/292,386 encounters) observed before the hurricane.

Mold-positive cultures from patients with invasive infections increased by 17.5% following Hurricane Harvey. Fungal species cultured from patients with invasive mold infections included: Aspergillus (44% of cultures post-Harvey; 43.8% of cultures pre-Harvey), Fusarium species (12.3% post-Harvey; 4.6% pre-Harvey), and Penicillium species (5.5% post-Harvey; 6.4% pre-Harvey). The presence of Aspergillus on culture also increased in patients who did not have invasive infections—from 29.2% pre-Harvey to 37.3% post-Harvey—potentially representing rises in colonization.

Most patients were hospitalized (95.1%), and almost one-quarter of patients with invasive mold infections died in the hospital (24.2%). Among hospitalized patients, people who received a fungal ICD-10 code had a significantly greater risk of death (adjusted odds ratio: 16.83). The mortality rate was 20% among proven cases, 19.5% among probable cases, and 29.1% among surveillance cases.

Demographic and clinical risk factors

The median age of patients with invasive mold infections was 57, 67.9% were male, and approximately 42% of cases were associated with the cancer care hospital. Almost half of patients (45.1%) required intensive care. The most common underlying conditions that placed patients at risk for invasive mold infection included: lymphopenia (58.2%), cancer (48.2%), diabetes (31.3%), and neutropenia (20.3%). The most common sites of infection were: pulmonary (60.7%); skin, tissue, or wounds (16.6%); and the sinuses (9.7%).

More than one-quarter of patients (25.8%) had neither risk factors for invasive mold infection nor clinical features of infections such as radiologic abnormalities, representing a rise in patients with non-traditional risk factors from the 19.2% observed before Hurricane Harvey.

Factors associated with a higher risk of in-hospital all-cause mortality included:

• Antifungal treatment in the 90 days before to the 60 days after the date of infection (13.83 adjusted odds ratio), though antifungal prophylaxis within the same time period was not associated with a greater risk of death.
• Injury and surgery: 8.3 adjusted odds ratio
• Central venous catheterization: 2.9 adjusted odds ratio
• Cancer: 1.93 adjusted odds ratio
• Lymphopenia: 1.74 adjusted odds ratio

Conclusions

The authors note that the study differed from previous analyses that did not find an increase in invasive mold infection incidence following Hurricanes Katrina, Rita, and Harvey. Because of the observed rise in invasive mold infection incidence following extreme flooding in Houston, the increase in cases among people who did not have traditional risk factors, and the ongoing effects of climate change, systems for epidemiological surveillance of fungal infection and colonization, along with an investigation into the role of emergency antifungal prophylaxis, are urgently needed in areas that are subject to extreme flooding.