The data of 25 patients with confirmed cases of HPS were analyzed (17 men, 8 women, mean age 33.4 years, range 15-63 years). The infection was most commonly acquired through farm or timber work (40%), and 76% of the patients were rural residents. The incubation period, estimated in 14 cases, was 9.8 + 7.5 days (range 3-28 days). Symptom duration before admission was 5 + 1.8 days (range 2-10 days); 12 (48%) of 25 patients had requested previous medical care on one to four occasions before admission; and in 20 (80%) of 25 case-patients, a diagnosis other than HPS was initially suggested.

The differential diagnoses before HPS was considered were community-acquired pneumonia or pneumonitis (including viral, interstitial, atypical, P. carinii pneumonia) (9 patients), sepsis (with/without shock or acute respiratory distress syndrome) (5 patients), fever syndrome (4 patients), influenza (3 patients), acute abdominal condition (1 patient), acute pyelonephritis (1 patient), acute tonsillitis (1 patient), bacterial meningitis (1 patient), typhoid fever (1 patient), acute diarrhea (1 patient), and myeloproliferative syndrome (1 patient). The main clinical signs and symptoms at admission were fever in 24 (96%) of 25 patients, myalgia in 24 (96%), asthenia in 19 (76%), headache in 15 (60%), abnormal breathing sounds on auscultation in 15 (60%), abdominal pain in 13 (56%), anorexia in 12 (48%), dry cough in 10 (40%), chills in 8 (32%), vomiting or nausea in 6 (24%), cyanosis in 5 (20%), diarrhea in 2 (8%), hemoptoic sputum in 2 (8%), and epistaxis, metrorrhagia, generalized maculopapular rash, consciousness alteration, lumbar pain, and odynophagia, each in 1 patient. Pseudoinfluenza syndrome was found in 17 (68%) patients. Bleeding was documented in 16 (64%) of 25 patients, and 3 (12%) patients required medical care for this reason. Tables 1 and 2 show the type and severity of bleeding manifestations and their relationship to platelet count.

The results of laboratory tests at admission can be seen in Table 3; thrombocytopenia was indicated by a platelet count of <100,000 in 23 (92%) case-patients. The mean hematocrit values were 49.8 + 6.4, and three patients showed a value <45%. Twelve (48%) patients showed alterations in renal function, six (24%) exhibited a creatinine level >2.0, and one patient required hemodialysis. In 18 (72%) patients, liver function tests showed alterations, but only one patient had liver failure.

Hypotension at admission was observed in 12 (48%) of 25 patients, and 16 (60%) had hemodynamic instability (reaching 93% in the most severe group). Eight patients experienced refractory shock. In one patient, a Swan-Ganz catheter was used in a late stage of the infection, and the hemodynamic profile was consistent with septic shock, with low systemic venous resistence and high cardiac output. These conditions were also consistent with a case of nosocomial gram-negative sepsis; this was later demonstrated by blood culture. This case was the only one in which a bacterial overinfection was suspected and confirmed.

Radiographic changes exhibited at admission consisted of bilateral interstitial infiltrates in 14 patients (56%), alveolar infiltrates in 4 (16%), and mixed infiltrates in 5 (20%) patients. In two (8%) patients, the disease progressed without pulmonary infiltrates. The pulmonary infiltrates, when present, were bilateral and extended, involving 4 or 5 lobules in all cases. In seven (28%) patients, radiographic infiltrates were found to progress after 48 hours. Two patients showed pleural effusion. Samples from two patients with no pulmonary involvement underwent serologic tests because the patients lived in a rural region where previous cases had been found and because they had a clinical history of fever, myalgia, and malaise. One patient had thrombocytopenia (platelet count 53,000), and both had raised hematocrit values (50% and 46%).

The mean PAFI at admission was 160 + 121 (range 40-508). Eight (32%) patients received oxygen either nasally or by mask, and in three (12%), supplemental oxygen was not required. Fifteen (60%) patients were admitted to the ICU, and 14 (56%) received mechanical ventilation early, 9 (64%) of 14 during the first 24 hours after admission, 4 before 48 hours, and 1 patient on day 3. The mean period of mechanical ventilation was 4 days (range 1-13 days). Ten (40%) patients died, seven before completing 1 day in the hospital, and one patient died later of septicemia caused by gram-negative bacteria. All patients who died showed diffuse and rapidly progressive interstitial pulmonary infiltrates, compatible with massive pulmonary edema, severe respiratory failure and refractory shock. Finally, electromechanical dissociation and asystolia were noted.

For the 15 patients admitted to the ICU, the mean APACHE II score was 18.3+8.7. No patient with APACHE II <12 died, and the mortality rate for the patients with APACHE II >12 was 77% (10/13) (p = 0.000). In the group of 10 case-patients who received methylprednisolone, 2 (20%) died, which contrasts with 8 (53%) deaths of 15 who did not receive this drug (p = 0.21).

We administered prednisolone to 10 patients; the decision to use it was made on admission for half of the case-patients and 2 to 4 days later for the remaining case-patients. One patient had hyperglycemia associated with the use of methylprednisolone. Nineteen (76 %) patients received antibiotic drugs, which were usually discontinued after confirmation of the viral cause. The variables with statistically significant differences between survivors versus nonsurvivors are shown in Table 4. Other variables associated with death were cyanosis (p = 0.005), renal function alteration (p = 0.001), shock (p = 0.000), ICU admission (p = 0.001), and mechanical ventilation (p = 0.001).

Statistically significant differences were not observed for the following variables: hematocrit, platelet count, number of symptomatic days before admission, PaO₂, leukocyte count, plasma sodium and potassium, bilirubin, glutamic-oxaloacetic transaminase (GOT) and glutamic-pyruvic transaminase (GPT), type and number of bleeding episodes, alteration in consciousness, growth of 50% in radiographic images during the first 48 hours, and the use of methylprednisolone.

We identified three groups of case-patients. The first included two case-patients with only prodromic symptoms without pulmonary involvement. The second group (n = 9) consisted of patients who had a self-limited form of the disease, usually with different degrees of interstitial pulmonary infiltrates, who did not require intubation and received oxygen through noninvasive methods (including one case-patient who needed no oxygen), were hemodynamically stable, and had a low APACHE II (<12). The third group (n = 14) had severe forms of the disease with a high mortality rate (APACHE II >12), required intubation and mechanical ventilation, with hemodynamic instability in 13 (93%) and refractory shock in 8 (57%). Pulmonary infiltrates were usually alveolar. The mean hospital stay was 13.6 days + 24.7, ranging from hours for those patients who died quickly to 120 days in one patient with a peridural hematoma and severe secondary paraparesia as a consequence of a lumbar puncture.

Our study confirms the existence of different clinical forms of hantavirus disease (7), including some without pulmonary involvement. A higher mortality rate was clearly associated with the most severe form (APACHE >12) of the disease, and death frequently occurred within the first 24 hours of hospital admission (70%). The main variables associated with death were APACHE II score, low PAFI, cyanosis, polypnea and tachycardia at admission, alterations in renal function, shock, and mechanical ventilation.

The incubation period was estimated for 14 cases and ranged from 3 to 28 days (x = 9.8 + 7.5), which is consistent with other reports (7,14,15). The symptoms at admission were similar to those described in case-patients in the United States (16,17) and Chile (10,14). Fundamentally, the initial clinical presentation is similar to influenza without rhinorrhea, high fever, myalgia, headache, or asthenia, and often accompanied by abdominal symptoms such as pain, nausea, vomiting, and diarrhea. After a prodromal period of 2 to 7 days, respiratory symptoms appear with dyspnea, respiratory failure, and hemodynamic instability. Chest x-rays show bilateral interstitial infiltrates of different degrees of rapid progression.

Nevertheless, the infection can be manifested in other ways, and many cases do not go beyond the prodromal stage or elicit any symptoms. In January 1997, the first hantavirus serologic study was carried out in Chile; of 64 serum samples collected in a rural location, 7 were positive for specific Andes virus nucleoprotein IgG. These 64 samples represented almost the whole population living in the area of the first Chilean case-patient’s residence. Thus nearly 10% of this rural population had had contact with the virus, although none of the persons who provided a sample remembered a severe illness (7).

From 1995 to 2000, a total of 10 case-patients with mild forms of hantavirus disease were detected in Chile. The disease progressed in these case-patients without major pulmonary involvement, but the available information does not allow a close analysis of the clinical features of these cases, since most findings were from epidemiologic and serologic investigation (18). We found two case-patients who were admitted to the hospital because of a clinical course of fever, myalgia, and malaise, whose disease progressed with no pulmonary involvement as indicated by serial x-rays during their hospital stay. The diagnosis was suggested after considering the epidemiologic data (they lived in a rural area from which other cases had come) and suggestive laboratory test results: thrombocytopenia indicated by a platelet count of 53,000 in one and raised hematocrit value in both (50% and 46%). In these patients, the symptoms could have been confounded with a viral disease such as influenza and thereby been underdiagnosed because of the nonspecific character of the symptoms. In 80% of our case-patients, the initial diagnosis was incorrect, and 48% had received medical care on one to four occasions before hospital admission. By asking epidemiologically oriented questions in these endemic countries and regions, physicians may be able to make an earlier diagnosis of HPS.

In the series of patients described by Duchin et al. (16) in the United States, the early symptoms were similar to those described in the Asian and European forms of hantavirus disease known as hemorrhagic fever with renal syndrome. Nevertheless, no patient showed hemorrhages, renal involvement was minimal, and neither oliguria nor renal failure was observed in any case. The 1995 outbreak in Argentina was described as similar but details were not given (19). By contrast, the cases described in Chile are different, given the hemorrhagic manifestations observed. In Coyhaique (Chile), Tapia (10) saw petechiae in 38% (9/24), epistaxis in three pediatric patients and microhematuria in 16 patients (66.6%). In Region IX of Chile, Castillo (14,15) described a higher frequency of bleeding events (71% and 81%), including hematemesis, hemoptisis, epistaxis, and puncture site bleeding, similar to what we observed in 64% of the patients (Tables 1 and 2). In three patients of our series, a bleeding event was the reason for seeking healthcare. Thus, hemorrhagic manifestations appear to occur frequently in cases of hantavirus in Chile.

Concerning renal function, Tapia (10) observed raised creatinine of >1.5 in 7 of 24 cases, which did not correlate with the severity of clinical course. In this study, acute renal failure was considered a terminal phenomenon. Castillo reported two case-patients with acute renal failure; one required hemodialysis, and creatinine values were elevated (1.2-3.3 mg/dL) in 54% of the patients (14,15). In our series, 48% of the patients fulfilled the American Thoracic Society criteria for altered renal function (9), and the difference between the creatinine values of nonsurvivors and survivors measured at admission (p = 0.0003) was statistically significant. In 24%, creatinine values were >2.0, and one patient required hemodialysis. We believe that the use of common criteria to define renal failure in larger series can help estimate the changes in renal function, which seem to be more important than previously described and have a prognostic value. With regard to laboratory tests, raised hematocrit values and low platelet counts are exhibited by almost all patients at admission, and thus are important diagnostic elements, although they have no prognostic value.

Corticosteroids have the potential of modulating intrapulmonary inflammatory response by modifying the proinflammatory cytokine levels such as interleukin-1-b and tumor necrosis factor-µ (TNF-µ), and have been used in the management of severe pulmonary disorders, with both an infective as well as a noninfective etiology such as miliary tuberculosis, Pneumocystis carinii pneumonia, vasculitides, and gastric acid aspiration (20).

In our study, 2 of 10 patients who received methylprednisolone (20%) died versus. 8 of 15 (53%) in the group who did not receive this drug, although this difference was not statistically significant (p = 0.21). In larger series, a significant effect could perhaps be demonstrated. When the first cases were being diagnosed, however, we did not consider the use of corticosteroids; no protocol on how to use them existed, and most severely ill patients died very quickly and thus did not receive corticosteroids; their usage probably would not have changed the fatal course of the disease in this group, however.

The mortality rate in our series was directly related to the type of clinical presentation. In the severe form, with APACHE II >12, the mortality rate reached 77%, in contrast to those with a score <12, among whom no deaths were observed. From 1993 to July 2000, a total of 123 confirmed HPS cases were reported in Chile, 61 (49.6%) of which were fatal. This rate has decreased during the last 2 years, partly because of the improvement in diagnostic capacities and greater knowledge of the disease, which allows quicker identification and more effective treatment (18), but we think that this reduction also may be due to a dilution effect because more mild cases are found.

In our series, we distinguish three groups of patients. The first (n = 2), probably the least known, corresponds to patients with only prodromal symptoms, especially a pseudoinfluenza course with no pulmonary involvement, who are often not admitted to a hospital and are not recognized to have HPS. Correct diagnosis of these cases improves if one makes a good epidemiologic questionnaire. Groups 2 and 3 allow the patients to be classified according to clinical and radiologic criteria, severity, treatment, and prognosis. Use of uniform criteria is required to classify the patients infected by hantavirus with different clinical presentations, to thereby comparatively evaluate the mortality rate and the effectiveness of therapeutic measures.