Results
Data Synthesis
Our search strategy permitted to identify 821 papers, of which 311 were excluded by title and abstract evaluation. Full-text papers were then assessed for eligibility according to the criteria outlined above. Among the 213 papers included, some were small case series, eventually the number of cases analyzed was 244 (Figure 1).
Countries
Reports from highly endemic countries were 65/244 (27%), with India Argentina), Brazil and Peru accounting for more than two thirds. Only four cases were reported from the whole of Africa, three of which in South Africa, a state where adequate diagnostic facilities are available. We collected 83/244 (34%) reports from North America (USA and Canada), 58/244 (24%) from Europe (Belgium, France, Germany, Greece, Italy, the Netherlands, Spain, Switzerland, UK) and five (2%) from Oceania (Australia and New Zealand). In these areas of low/no endemicity, half of the patients were immigrants (70/146, 48%), while a few subjects were veterans (5/146, 3%) who presumably acquired the infection during military service in an endemic country. Other areas of low endemicity where cases have been reported are in Eastern Asia (21 cases, mostly from Japan and Taiwan), the Arabian peninsula (nine cases, mostly from Kuwait and Qatar and Israel (three cases). Countries such as Iran Turkey and Venezuela that might be presumed at medium to high prevalence, account for only one case each.
Triggers for Development of HS/DS
According to the case definitions, 171 cases were classified as hyper infection and 73 cases as dissemination.
A high percentage of patients (67%: 164/244) were under corticosteroids: most of them presented clinical conditions causing immune suppression per-se or due to other related therapies (for instance leukemia, rheumatic conditions, transplant), as it is shown in Table 1 . On the other hand, a few patients were taking steroids for eosinophilia and/or a specific symptoms caused by S. stercoralis itself (data reported in Table 1 too). A patient even underwent bone marrow transplant because of an unexplained eosinophilia misdiagnosed as "idiopathic hypereosinophilic syndrome"; after receiving steroids and immunosuppressive therapy he developed HS (but only limited autopsy was performed, so we cannot rule out DS) and died.
Transplant is surely an event that poses the Strongyloides - infected patient at high risk of developing HS/DS. We collected 28/244 (11.5%) cases of HS/DS in transplant patients, of whom 19 (68%) died. A couple of patients who developed hyper infection also had co infection with CMV. All the surviving patients received ivermectin, either as single treatment (1 patient) or in combination with albendazole (7) or thiabendazole (1).
HTLV-1 infection is a well known risk factor (sometimes in association with related haematological malignancies), of which we found 24/244 (10%) reports. Ten of the 24 patients (42%) died. One patient had HTLV-1-HIV co infection; he developed an E. coli meningitis but successfully responded to ivermectin, two doses given some days (not specified how many) apart.
We found 38/244 (15%) reports on HIV-positive patients, 26 (68%) of whom died (Figure 2). Seven HIV patients were also receiving steroids for suspected Pneumocystis jiroveci pneumonia, immune reconstitution inflammatory syndrome, misdiagnosis of asthma, Wegener granulomatosis, toxoplasmosis encephalitis, cerebral TB with vasculitis; six of them died.
A few reports/case series describe severe strongyloidiasis in patients with alcoholism and malnutrition. An apparently immunocompetent patient developed hyper infection and died two days after having started therapy with thiabendazole. Unfortunately autopsy was denied.
Diagnosis
Eosinophilia was present in 55/244 cases (22.5%) overall, and only in 12/73 cases (16.4%) of dissemination. In all cases S. stercoralis was found at microscopy examination of biological samples. Serology was performed only in 16/244 patients (6.5%) (Table 2). In a couple of organ transplant recipients, an ELISA test was negative pre-transplant, but resulted positive in the deceased donors (test performed retrospectively). In other two cases serology (ELISA) was negative: a HIV-infected person, who had larvae in stool and sputum and a patient with dermatomyositis, under chronic treatment with prednisone and methotrexate, who died from disseminated strongyloidiasis (larvae found at autopsy in skin, lungs, small and large bowel, gall bladder, vessels of meninges and cervical spinal cord).
Diagnosis was obtained post mortem in 29 cases (12%).
Therapy
Therapies given were very different in relation to the drugs used and the length of treatment.
In Table 3 we summarize the drugs used. In the "other drugs" group we found mebendazole, cambendazole, levamisole, pyrantel pamoate, diethylcarbamazine.
Albendazole was used as a single drug even in recent case reports; since 2008 we found patients treated with albendazole only in reports from Pakistan, Romania, Taiwan, Israel, Kuwait, Argentina, Malaysia, Greece, Thailand.
In most cases the administration route was oral, but due to severe clinical conditions of patients, administration via nasogastric tube, subcutaneous injection (veterinary formulation) and retention enema were used, too.
A patient who developed disseminated strongyloidiasis after an empiric steroid treatment for pruritic rash was treated with albendazole. Only one dose could be given, as the patient died. After his death, a review of his clinical records showed that he had been previously diagnosed with strongyloidiasis and treated with a 3-day course of albendazole; although serology persisted positive and eosinophilia was still present 6 and 12 months after treatment, the patient did not receive any further therapy. Another patient who died from Strongyloides hyper infection had never been treated previously, despite a positive serology.
Outcome
The recorded deaths were 153/244 (62.7%). A similar fatality rate was observed in patients with dissemination (50/73 = 68.5%) and with hyperinfection (102/171 = 60%).
All 42 of 244 patients who did not receive any therapy died. Excluding patients treated with combination therapy, we observe that 25/34 (73%) patients treated with albendazole died, while deaths among patients treated with ivermectin and thiabendazole were 18/38 (47%) and 28/55 (51%), respectively.