As it is unclear to which extent an outpatient regimen of current standard immunosuppression for autoimmune disease may confer a risk for HSV-1 pneumonitis, we retrospectively reviewed the charts of ~1400 patients with rheumatoid arthritis (~1000), ANCA-associated vasculitis (~80), and systemic lupus erythematosus (SLE) (~320) that were followed during the years 2000–2007 in the nephrology and rheumatology outpatient clinic at the Medical Policlinic, LMU University of Munich. During the study period 766 ward-admissions of patients with rheumatologic/vasculitic/connective tissue autoimmune disease occurred. Among those there were 63 (8%) episodes of admission from 63 outpatients of our clinic presenting with respiratory deterioration that were ultimately diagnosed to be due to pneumonia or pneumonitis (Table ). As a rough estimate one can infer an incidence of community-acquired respiratory infection or pneumonitis requiring hospitalization in outpatients with autoimmune disease of [63 patients * 100/(1400 patients *8 years)], i.e., ~0.5% per year.
| Table 1Classification and clinical course of the study population, i.e., 63 patients with rheumatic/vasculitic or connective tissue disease admitted to hospital and diagnosed to have pneumonia/pneumonitis |
In 23 of these 63 patients bronchoscopy with lavage was performed for diagnostic purposes within the first four days of admission, which in 14 (61%) also comprised PCR analysis for HSV-1. Thereby a total of six patients (10% of admissions for pneumonia) with positive HSV-1 DNA detection in alveolar fluids was identified. In order to estimate the degree of immunosuppressive strength in the patient population, an arbitrary score ranging form 0–4 was applied (see methods).
Relevant clinical, imaging and laboratory data on the six patients with a positive HSV-1 DNA detection on BAL are summarized in Table .
| Table 2Synopsis of patient data on all cases with a positive HSV-1 result on PCR of bronchoalveolar lavage |
Two cases (Case 1 and 2) in which no other viral, bacterial or fungal agent than HSV-1 on BAL PCR was found and in which imaging studies were suggestive for HSV-1 pneumonitis were successfully treated with acyclovir leading to rapid clinical improvement. In both cases, the initial chest radiograph and thoracic CT showed diffuse infiltrations or ground glass opacities. Representative images are shown in Figure . HSV-1 infection may feature a macroscopically vulnerable aspect of the bronchial mucosa as a non-specific sign (Figure ). In order to increase diagnostic certainty in Case 2, immunohistochemical staining of bronchioalveolar cells for HSV-1 was performed, which proved positive (Figure ). In both cases inclusion bodies of epithelial cells from BAL were present. Our diagnosis in Case 1 and 2 was isolated HSV-1 pneumonitis, which was supported by the response to acyclovir in the absence of prolonged antibacterial treatment.
Case 3 and 4 displayed an initial co-infection with fungal or bacterial agents and HSV-1 and required prolonged mechanical ventilation. Imaging studies were compatible with viral pneumonitis, but in Case 4 infiltrates suggesting bronchopneumonia were also clearly present. The mucosa was macroscopically inflamed in both cases. While Case 3 had inclusion bodies on cytology, such studies were not pursued in Case 4 due to obvious bronchopneumonia. Both patients had a fatal outcome after a prolonged in-hospital stay despite adequate antiinfective treatment that included timely administration of acyclovir.
Cases 5 and 6 are different from the former four cases therein that no bilateral groundglass opacities were present on thoracic CT and no antiviral treatment was initiated. In Case 5, a high HSV-1 viral load from BAL fluids was ignored given the detection of Influenza A virus with S. aureus and Ps. aeruginosa. The patient improved on antibacterial treatment alone and was discharged. He was readmitted one month later with de-novo respiratory deterioration that resulted in the need for intubation after the administration of a series of antimicrobial courses that always excluded acyclovir. CT and a second BAL now showed ground glass opacities/ARDS and an even higher HSV-1 viral load. The patient died from ARDS before antiviral treatment could be initiated. Case 6 featured an unilateral bronchopneumonic infiltrate and was treated with standard antibacterial regimens until Mycobacterium kansasii grew from the initial BAL. Treatment was changed accordingly and the patient recovered without the administration of antivirals.
In an area of controversy, the first two cases provide further evidence that in selected patients, if ascertained to be the sole agent, HSV-1 is a relevant pathogen of pneumonitis that is amenable to treatment. Cases 3–6 reflect the great difficulty of dissecting the role of HSV-1 in infections with multiple pathogens [
19] and illustrate that HSV-1 pneumonitis is a diagnosis of exclusion requiring a complete microbiological workup. It remains speculative if timely suppression of HSV-1 viral loads by antiviral treatment in Case 5 would have altered the patient's course.
In the present study of 63 patients with autoimmune disease and pneumonia/pneumontis, no offending agent could be found in 60% of cases. In the remaining 40% a broad spectrum of bacteria, fungi and viruses was identified (Figure ). Of note, one case of pneumonitis was due to methotrexate. Two cases (3.2%) were most likely due to HSV-1, while HSV-1 was a bystander of other agents in the remaining four cases.
When comparing the six HSV-1-positive cases to those where HSV-1 was neither detected nor suspected with respect to the severity of immunosuppression, a significantly higher degree of immunosuppression in the HSV-1-positive group was noted (p < 0.01, two-sided Mann-Whitney analysis) (Figure ). This analysis was performed including 56 of the 63 patients where reliable information on maintenance immunosuppression was available. In order to exclude selection bias, the comparison was also performed using the subgroup of 8 patients that were proven to be HSV-1 negative by BAL. Albeit not reaching the same level of significance, a similar trend was observed with a similar median immunosuppression severity score of 2.0 in the HSV-1 negative group (data not shown). Though great caution must be applied in interpreting small numbers, it is intriguing that patients with Wegener's Granulomatosis and vasculitis, who had the strongest immunosuppression (score 3.3–3.5) also had a HSV-1 positive BAL more often (15–25% of pneumonia/pneumonitis) than Rheumatoid arthritis and SLE patients (6–8% of pneumonia/pneumonitis) who were also on a milder immunosuppressive regimen (score 1.5–1.6) (Table ).