ABSTRACT.
Human strongyloidiasis is a potentially life-threatening parasitic disease among immunocompromised hosts. We aim to determine the factors and mortality associated with disseminated strongyloidiasis. We conducted a U.S.-based multicenter retrospective cohort study to determine 90-day clinical outcomes for people diagnosed with Strongyloides infection in the TriNetX patient database. We identified adult patients with the International Classification of Diseases (10th revision, clinical modification) code for Strongyloides infection (B78) or a positive Strongyloides IgG antibody test and captured outcomes at 90 days. We identified 5,434 patients with strongyloidiasis, of whom 48 had disseminated strongyloidiasis for 0.9% prevalence of disseminated disease. Systemic connective tissue disorders, pulmonary eosinophilia, liver cirrhosis, blood disorders (monoclonal gammopathy, aplastic anemia, and lymphoid malignancy), malnutrition, alcohol use disorder, and transplantation status were frequent in patients with disseminated disease. Mortality was significantly higher in people with disseminated disease at 30 days (21%). The 90-day risk of hospitalization, bacteremia, and acute respiratory distress syndrome (ARDS) was higher in those with disseminated infection. People with disseminated strongyloidiasis had a heightened risk of hospitalization, bacteremia, acute respiratory distress syndrome, and mortality. The population at risk for severe strongyloidiasis infection is evolving, reflecting conditions in which glucocorticoids or additional immunosuppressive medications are commonly used for treatment.
INTRODUCTION
Human strongyloidiasis is a potentially life-threatening parasitic disease caused by the nematode Strongyloides stercoralis.1 Strongyloides hyperinfection syndrome (SHS) occurs when many migrating larvae overwhelm the body’s immune response. Infection with human T-lymphotropic virus type 1 (HTLV-1) and glucocorticoid use have been identified as the most critical risk factors for hyperinfection. The related entity of disseminated strongyloidiasis occurs when the larvae infect numerous organs. Both hyperinfection and dissemination are associated with high mortality in untreated patients. An increased population of immunosuppressed patients in the United States are potentially at risk for worse outcomes in the setting of Strongyloides infection. However, owing to the rarity of this condition, there is no accurate estimate of the risk of dissemination and mortality and descriptions of real-world risk factors for disseminated strongyloidiasis. Using a U.S.-based federated network, we aimed to determine the characteristics and mortality associated with disseminated strongyloidiasis.
MATERIALS AND METHODS
Study design and case definition.
We conducted a retrospective cohort study to determine 90-day clinical outcomes for people diagnosed with Strongyloides infection in the TriNetX patient database (https://trinetx.com). TriNetX is a global research network database well suited for studies of less frequently encountered infectious diseases.2,3 Contributing healthcare organizations, most often academic medical centers, deliver electronic medical record data to TriNetX, either structured or processed by natural language processing (eMethods, supplement). We queried U.S.-based healthcare institutions in the TriNetX database to identify adult (≥18 years) patients with the International Classification of Diseases (10th revision, clinical modification; ICD-10-CM) code for Strongyloides infection (B78) or a positive Strongyloides IgG antibody test (optical density unit value ≥1.1 U/mL) from database inception (approximately 2015–2016) to August 2023. We used ICD-10-CM code B78.7 to identify patients with disseminated strongyloidiasis when the larvae invade numerous organs not classically associated with the parasite’s lifecycle. The ICD-10-CM codes do not distinguish between SHS and disseminated strongyloidiasis.
Covariates and outcome measures.
We extracted data regarding patient demographics, diagnoses, comorbidities, procedures, laboratory testing, and medications (Supplemental Tables 1–4). For patients with multiple encounters with Strongyloides, we analyzed the index event or the earliest encounter with a Strongyloides diagnosis and captured the proportion of deaths as the primary outcome at 30 and 90 days. We also measured secondary outcomes of hospitalization, acute respiratory distress syndrome (ARDS), peritonitis, bacterial meningitis, and bacteremia within 90 days (Supplemental Table 4).
STATISTICAL ANALYSES
We use means and standard deviations for continuous variables and frequency and proportions for categorical variables for the bivariate analysis. We determined the prevalence of disseminated infection by dividing the number of people with ICD-10-CM code for disseminated strongyloidiasis by the total number of people with diagnosis code or laboratory evidence of Stronyloides infection. We determine the mortality odds ratio with 95% CIs. We compared continuous data using independent t tests and categorical data using chi-square or Fisher’s exact test. A P-value <0.05 was considered significant.
RESULTS
We identified 5,434 patients with strongyloidiasis, of whom 48 had disseminated strongyloidiasis. People with disseminated infection were older and more likely to have race identified as White or American Indian/Alaska Native (Table 1). Fever, malaise, dyspnea, and abdominal pain were more frequently present in those with disseminated infection. Wheezing, hemoptysis, and altered mental status were present in fewer than 10 people (20.8%) with disseminated disease. Nearly 60% of individuals with disseminated infection had sepsis, and half had chronic lower respiratory conditions or a history of cancer. Systemic connective tissue disorders, pulmonary eosinophilia, liver cirrhosis, blood disorders (monoclonal gammopathy, aplastic anemia, and lymphoid malignancy), malnutrition, alcohol use disorder, and transplantation status were also more frequent in disseminated versus nondisseminated disease. HIV and HTLV-1 were present in fewer than 10 patients. Laboratory findings were remarkable for lower hemoglobin and elevated lactate dehydrogenase in those with disseminated infection. The mean eosinophil count was elevated in both disseminated and nondisseminated disease (5.9 ± 9.8 × 103/µL and 7 ± 9.1 × 103/µL, respectively; P = 0.473). Glucocorticoids, immunosuppressants, and chemotherapy were more frequent in people with disseminated disease.
Table 1.
Clinical features of patients with strongyloidiasis by dissemination status
| Characteristic Name* | Disseminated† | Nondisseminated† | |
|---|---|---|---|
| Mean ± SD, n (%) | N = 48 | N = 5,386 | P-Value |
| Demographics | |||
| Age at index diagnosis | 56.2 ± 15.6 | 49.4 ± 19.5 | 0.017 |
| Sex (assigned at birth) | |||
| Male | 29 (60.4%) | 3,108 (57.7%) | 0.699 |
| Race/ethnicity | |||
| White | 26 (54.2%) | 2,147 (39.8%) | 0.043 |
| Hispanic or Latino | 14 (29.2%) | 1,242 (23%) | 0.316 |
| Black | ≤10 | 1,023 (19%) | 0.952 |
| Asian | ≤10 | 721 (13.4%) | 0.131 |
| American Indian or Alaska Native | ≤10 | 36 (0.7%) | <0.0001 |
| Symptoms | |||
| Dyspnea | 29 (60.4%) | 1,245 (23.1%) | <0.0001 |
| Malaise and fatigue | 25 (52.1%) | 1,095 (20.3%) | <0.0001 |
| Fever | 22 (45.8%) | 699 (13%) | <0.0001 |
| Diarrhea | 21 (43.8%) | 729 (13.5%) | <0.0001 |
| Abdominal and pelvic pain | 20 (41.7%) | 1,543 (28.6%) | 0.047 |
| Sore throat | 20 (41.7%) | 1,237 (22.9%) | 0.002 |
| Cough | 19 (39.6%) | 978 (18.1%) | <0.0001 |
| Nausea and vomiting | 19 (39.6%) | 847 (15.7%) | <0.0001 |
| Headache | 15 (31.3%) | 652 (12.1%) | <0.0001 |
| Rash | 11 (22.9%) | 544 (10.1%) | 0.003 |
| Comorbidities | |||
| Sepsis | 28 (58.3%) | 654 (12.1%) | <0.0001 |
| Chronic lower respiratory diseases | 24 (50%) | 987 (18.3%) | <0.0001 |
| Neoplasms | 24 (50%) | 1,451 (26.9%) | <0.0001 |
| Type 2 diabetes mellitus | 19 (39.6%) | 1,419 (26.3%) | 0.038 |
| Chronic kidney disease | 17 (35.4%) | 1,327 (24.6%) | 0.084 |
| Noninfective enteritis and colitis | 16 (33.3%) | 673 (12.5%) | <0.0001 |
| Pleural diseases | 16 (33.3%) | 648 (12%) | <0.0001 |
| Bacteremia | 15 (31.3%) | 336 (6.2%) | <0.0001 |
| Eosinophilia (ICD-D72.1) | 14 (30%) | 1,185 (22%) | 0.203 |
| Systemic connective tissue disorders | 12 (24%) | 215 (4%) | <0.0001 |
| Pulmonary edema | 11 (22.9%) | 292 (5.4%) | <0.0001 |
| Malnutrition | 11 (22.9%) | 588 (10.9%) | 0.008 |
| Liver fibrosis | ≤10 | 609 (11.3%) | 0.038 |
| COVID-19 | ≤10 | 54 (1%) | <0.0001 |
| Transplanted organ and tissue status | ≤10 | 519 (9.6%) | 0.009 |
| Alcohol-related disorders | ≤10 | 383 (7.1%) | <0.0001 |
| Alcoholic liver disease | ≤10 | 339 (6.3%) | <0.0001 |
| Monoclonal gammopathy | ≤10 | 33 (0.6%) | <0.0001 |
| Aplastic anemia | ≤10 | 240 (4.5%) | <0.0001 |
| Lymphoid tissue malignancy | ≤10 | 224 (4.2%) | <0.0001 |
| Neutropenia | ≤10 | 182 (3.4%) | <0.0001 |
| HIV | ≤10 | 158 (2.9%) | <0.0001 |
| HTLV-1 | ≤10 | 10 (0.2%) | <0.0001 |
| Pulmonary eosinophilia | ≤10 | 10 (0.2%) | <0.0001 |
| Stem cell transplant status | 0 (0%) | 17 (0.3%) | 0.697 |
| Laboratory | |||
| Leukocytes (103/µL) | 9.3 ± 5.8 | 9.2 ± 68.5 | 0.992 |
| Hemoglobin (mg/dL) | 10.3 ± 2.5 | 12.3 ± 2.5 | <0.0001 |
| Platelets (103/µL) | 246.8 ± 141.9 | 236.1 ± 110 | 0.542 |
| Lymphocytes (103/µL) | 2.3 ± 1.8 | 2.5 ± 1.3 | 0.277 |
| Eosinophils (103/µL) | 5.9 ± 9.8 | 7 ± 9.1 | 0.473 |
| Alanine aminotransferase (IU/mL) | 53.7 ± 136.1 | 36.1 ± 129 | 0.417 |
| Lactate dehydrogenase (IU/mL) | 775.2 ± 1,386.4 | 310.9 ± 342.6 | <0.0001 |
| Ferritin (µg/L) | 575 ± 685.2 | 609.4 ± 3,400.8 | 0.961 |
| C-reactive protein (mg/L) | 47.9 ± 59.8 | 41.5 ± 66.2 | 0.644 |
| Erythrocyte sedimentation rate (mm) | 38.7 ± 33 | 35.9 ± 31.8 | 0.695 |
| CD4 count (cells/µL) | 527 ± 307.7 | 979.4 ± 3,995.6 | 0.722 |
| Medications | |||
| Glucocorticoids | 40 (83.3%) | 2,151 (39.9%) | <0.0001 |
| Prednisone | 25 (52.1%) | 931 (17.3%) | <0.0001 |
| Methylprednisolone | 21 (43.8%) | 849 (15.7%) | <0.0001 |
| Dexamethasone | 17 (35.4%) | 839 (15.6%) | <0.0001 |
| Hydrocortisone | ≤10 | 544 (10.1%) | 0.014 |
| Prednisolone | ≤10 | 208 (3.9%) | <0.0001 |
| Immunosuppressants | 11 (22.9%) | 418 (7.8%) | <0.0001 |
| Tacrolimus | ≤10 | 286 (5.3%) | <0.0001 |
| Mycophenolate mofetil | ≤10 | 225 (4.2%) | <0.0001 |
| Mycophenolic acid | ≤10 | 135 (2.5%) | <0.0001 |
| Azathioprine | ≤10 | 70 (1.3%) | <0.0001 |
| TNF-alpha inhibitors | 0 (0%) | 43 (0.8%) | 0.534 |
| Chemotherapy | 13 (27.1%) | 469 (8.7%) | <0.0001 |
| Antiparasitics | |||
| Ivermectin | 23 (47.9%) | 1,394 (25.9%) | 0.001 |
| Albendazole | ≤10 | 284 (5.3%) | <0.0001 |
| Outcomes‡ | |||
| Mortality (30 days) | 10 (21%) | 108 (2%) | <0.0001 |
| Mortality (90 days) | 10 (21%) | 181 (3.4%) | <0.0001 |
| Hospitalization (90 days) | 36 (75%) | 3,024 (56.1%) | 0.011 |
| ARDS (90 days) | 12 (25%) | 345 (6.4%) | <0.0001 |
| Bacteremia (90 days) | 10 (21%) | 161 (3.0%) | <0.0001 |
| Peritonitis (90 days) | 10 (21%) | 81 (1.5%) | <0.0001 |
| Bacterial meningitis (90 days) | 10 (21%) | 11 (0.2%) | <0.0001 |
ARDS = acute respiratory distress syndrome; HTLV-1 = human T-lymphotropic virus 1; ICD = International Classification of Diseases; TNF = tumor necrosis factor.
Variables captured non–time bounded before diagnosis (index event).
A value of 10 is the TrinetX data obfuscation threshold.
Captured within 1–3 months after diagnosis (index event).
Mortality was significantly higher in people with disseminated disease at 30 days (21% versus 2%, P <0.0001; OR: 12.5, 95% CI: 6.1–25.6) and at 90 days (OR: 7.1, 95% CI: 3.5–14.6, P <0.0001). The 90-day risk of hospitalization (OR: 2.2, 95% CI: 1.2–4.2, P = 0.011), bacteremia (OR: 8.0, 95% CI: 3.9–16.3, P <0.0001), and ARDS (OR: 4.6, 95% CI: 2.4–8.9, P <0.0001) were all higher in those with disseminated infection. The 90-day mortality among those hospitalized within 3 months was 11% versus 1.5% for those nonhospitalized. Peritonitis and bacterial meningitis were also present more frequently in people with disseminated disease. Up to 42% and 20% of disseminated disease patients received ivermectin and albendazole within 90 days, respectively.
DISCUSSION
This multicenter U.S.-based network found an overall 90-day mortality of 21% for people with a disseminated Strongyloides infection. Approximately three-quarters of people required hospital admission, and a quarter developed bacteremia and ARDS within 90 days. We also found chronic lower respiratory diseases, cancer, inflammatory bowel disease, systemic connective tissue disorders, and the reception of glucocorticoids, immunosuppressants, or chemotherapy as the most common associations linked to disseminated strongyloidiasis.
Data on well-established risk factors associated with disease severity are scarce. A case series and literature review of SHS found glucocorticoids (84%) with a median dose of 40 mg/day, immunosuppressants (25%), chemotherapy (18%), autoimmune disease (25%), hematologic malignancy (20%), and HIV infection (11%) as the most common associations.4 Percentages were similar to the ones found in our study. Although glucocorticoids remain a prevalent established risk factor, they may unveil non-HIV, nontransplant immunosuppressive risk conditions such as cancer, inflammatory bowel disease, and connective tissue disorder that require glucocorticosteroids for treatment. Our study shows advanced HIV or AIDS and HTLV-1 infection may be less predominant contemporary risk factors in developed countries. However, they may still be commonly found in the global South.5 Causes of mortality in patients with SHS and disseminated infection are often related to bacterial sepsis and end-organ infections (meningitis, peritonitis, hepatosplenic microabscesses).6 Previous case series have reported mortality close to 80% among immunocompromised patients and intensive care mortality of 60%.4,7 The in-hospital mortality rate was similar to a recently reported rate of 8%.8 The discrepancy in mortality may be attributable to different study populations and years of the studies. The lower treatment rates may be attributed to a lack of treatment pathway capture in half of those patients.
Our results should be interpreted with several limitations. ICD-10-CM codes do not distinguish between SHS and disseminated disease; therefore, we could not assess potential differences in clinical outcomes between these two conditions. The retrospective nature of the follow-up cohort can introduce selection bias. We captured the diagnosis of invasive disseminated strongyloidiasis through an ICD code, which can be subject to code errors and capture patients not meeting diagnostic criteria. The low number of patients did not allow for multivariable propensity score matching to identify stronger predictors of severity. However, the advantages of this study are that it uniquely compares disseminated disease with nondisseminated disease in people with strongyloidiasis and provides a mortality estimate.
CONCLUSION
Mortality, hospitalization, and complications (bacteremia, ARDS) of disseminated strongyloidiasis are remarkable. The population at risk for severe strongyloidiasis infection is evolving, reflecting conditions in which glucocorticoids or additional immunosuppressive medications are commonly used for treatment. Symptoms are nonspecific, and eosinophilia may not be reliable in identifying at-risk patients.
Supplemental Materials
Note: Supplemental materials appear at www.ajtmh.org.
Data Availability
The corresponding author had full access to data in the study and was ultimately responsible for submitting the manuscript for publication. The aggregated datasets generated and analyzed in the current study are available from the TriNetX platform. Any data displayed on the TriNetX platform in aggregate form, or any patient-level data provided in a data set generated by the TriNetX platform, only contains de-identified data as per the de-identification standard defined in Section 164.514(a) of the HIPAA Privacy Rule.
REFERENCES
- 1. Nutman TB, 2017. Human infection with Strongyloides stercoralis and other related Strongyloides species. Parasitology 144: 263–273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Vargas Barahona L. et al. , 2023. Previous corticosteroid exposure associates with an increased Pneumocystis jirovecii pneumonia mortality among HIV-negative patients: A global research network with a follow-up multicenter case–control study. Ther Adv Infect Dis 10: 20499361231159481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Henao-Martínez AF, Orkin CM, Titanji BK, Rodriguez-Morales AJ, Salinas JL, Franco-Paredes C, Tuells J, Chastain DB, 2023. Hospitalization risk among patients with mpox infection – A propensity score matched analysis. Ther Adv Infect Dis 10: 20499361231196683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Geri G, Rabbat A, Mayaux J, Zafrani L, Chalumeau-Lemoine L, Guidet B, Azoulay E, Pène F, 2015. Strongyloides stercoralis hyperinfection syndrome: A case series and a review of the literature. Infection 43: 691–698. [DOI] [PubMed] [Google Scholar]
- 5. Ye L, Taylor GP, Rosadas C, 2022. Human T-cell lymphotropic virus type 1 and Strongyloides stercoralis co-infection: A systematic review and meta-analysis. Front Med (Lausanne) 9: 832430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Marcos LA, Terashima A, Canales M, Gotuzzo E, 2011. Update on strongyloidiasis in the immunocompromised host. Curr Infect Dis Rep 13: 35–46. [DOI] [PubMed] [Google Scholar]
- 7. Igra-Siegman Y, Kapila R, Sen P, Kaminski ZC, Louria DB, 1981. Syndrome of hyperinfection with Strongyloides stercoralis. Rev Infect Dis 3: 397–407. [DOI] [PubMed] [Google Scholar]
- 8. Inagaki K, Bradbury RS, Hobbs CV, 2022. Hospitalizations associated with strongyloidiasis in the United States, 2003–2018. Clin Infect Dis 75: 1548–1555. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The corresponding author had full access to data in the study and was ultimately responsible for submitting the manuscript for publication. The aggregated datasets generated and analyzed in the current study are available from the TriNetX platform. Any data displayed on the TriNetX platform in aggregate form, or any patient-level data provided in a data set generated by the TriNetX platform, only contains de-identified data as per the de-identification standard defined in Section 164.514(a) of the HIPAA Privacy Rule.