Volume 4, Issue 11 pp. 1883-1888
Free Access

A Seroprevalence Study of West Nile Virus Infection in Solid Organ Transplant Recipients

Deepali Kumar

Corresponding Author

Deepali Kumar

The Immunocompromised Host Infection Service, Multi-organ Transplant and Infectious Diseases, University of Toronto, Toronto, Ontario, Canada, M5G 2N2,

*Corresponding author: Deepali Kumar, [email protected]Search for more papers by this author
Michael A. Drebot

Michael A. Drebot

Zoonotic Diseases and Special Pathogens, National Microbiology Laboratory, Health Canada, Winnipeg, Manitoba, R3E 3R2,

Search for more papers by this author
Susan J. Wong

Susan J. Wong

Wadsworth Center, New York State Department of Health, P.O. Box 22002, Albany, NY 12201-2002,

Search for more papers by this author
Gillian Lim

Gillian Lim

Foodborne, Waterborne and Zoonotic Infections Division, Centre for Infectious Disease Prevention and Control, Health Canada, Guelph, Ontario, Canada N1G 5B2, and

Search for more papers by this author
Harvey Artsob

Harvey Artsob

Zoonotic Diseases and Special Pathogens, National Microbiology Laboratory, Health Canada, Winnipeg, Manitoba, R3E 3R2,

Search for more papers by this author
Peter Buck

Peter Buck

Foodborne, Waterborne and Zoonotic Infections Division, Centre for Infectious Disease Prevention and Control, Health Canada, Ottawa, Ontario, Canada K1A 0L2

Search for more papers by this author
Atul Humar

Atul Humar

The Immunocompromised Host Infection Service, Multi-organ Transplant and Infectious Diseases, University of Toronto, Toronto, Ontario, Canada, M5G 2N2,

Search for more papers by this author
First published: 16 July 2004
Citations: 90

Abstract

West Nile virus (WNV) causes severe neurological disease in less than 1% of infections. However, meningoencephalitis may be more common in immunosuppressed transplant patients. In 2002, a WNV outbreak occurred in our region. To determine the spectrum of disease of community acquired WNV infection and assess public health behavior patterns in transplant recipients, we carried out a seroprevalence study. Patients were enrolled from outpatient transplant clinics in October 2002 and sera were screened for WNV. Questionnaires about WNV were provided to patients. Eight hundred sixteen organ transplant patients were enrolled. The seroprevalence of WNV IgM was 2/816 (0.25%; 95% CI 0.03–0.88%). By extrapolation to our entire transplant population of 2360 patients, and using data from hospital-based surveillance, the risk of meningoencephalitis in a transplant patient infected with WNV is estimated to be 40% (95% CI 16–80%). With regards to knowledge and behavior, 56% patients knew of and 47% used at least one protective measure against WNV. Only 33% used insect repellent. The risk of meningoencephalitis in transplant recipients is much higher than in the general population. There is incomplete knowledge and poor rates of compliance amongst patients with regards to WNV prevention.

Introduction

West Nile virus (WNV) is a single stranded RNA virus transmitted to human from infected mosquitoes (1,2). The virus first appeared in North America in a 1999 New York City outbreak and then rapidly spread throughout much of the United States and Canada resulting in large outbreaks of meningoencephalitis, particularly in 2002 and 2003 (3–6). In Canada, human cases of West Nile virus were first reported in the 2002 epidemic. In total, Canada reported over 400 cases of West Nile virus disease in 2002, with the majority in the Toronto and surrounding area. Significant media attention and public health resources were diverted to raise public health awareness.

Seroprevalence data obtained from West Nile virus outbreaks suggest that in previously healthy individuals, the majority of infections are asymptomatic (7,8). About 20% of infected people will have a self-limited febrile illness. It has been estimated that the rate of meningitis or encephalitis in the normal immunocompetent population is approximately one in every 140 infections or less than 1% (7).

Several unique findings were also noted during the 2002 WNV epidemic including the first documented cases of person-to-person WNV transmission through organ transplantation, blood transfusion, breastfeeding, and intrauterine transmission from mother to fetus (9–12). In one case, an infected organ donor transmitted West Nile virus infection to four transplant recipients (9). In addition to acquiring infection through the donor organ or blood transfusion, transplant patients may get community-acquired exposure from the bite of an infected mosquito. Because the Toronto area is home to a single large multi-organ transplant program, with over 2000 transplant recipients residing in this geographical location, we initiated specific educational initiatives to help prevent community-acquired infection in our patient population. Despite these measures, four cases of WNV meningoencephalitis occurred in our transplant population (13). To help determine the extent of serious, mild and asymptomatic infections in this population, and to assess knowledge and behavior patterns, we carried out an outpatient seroprevalence study in our transplant population.

Methods

The peak period of the West Nile virus epidemic in the Toronto and surrounding area was recognized as being from the end of August and beginning of September 2002. Prior to 2002, human cases of West Nile virus disease contracted within Canada had not been reported. This region is home to a single large multi-organ transplant program, which follows over 2000 transplant patients. Patient samples were obtained from October 1 to October 31, 2002, or approximately 4 weeks after the peak of the epidemic. This was after the first frost and mosquito activity had decreased markedly by this time. Patients attending routine transplant clinics were approached for enrollment in the study. To ensure as unbiased a sample as possible, only outpatients attending routine follow-up transplant clinics were approached. Inpatients, and patients attending ‘drop-in’ clinics for acute problems were not included. Any patient with suspected or known West Nile infection was not eligible for inclusion in the seroprevalence study. Patients living in a defined geographical area encompassing Toronto and the surrounding regions, during the summer and fall of 2002 were eligible for inclusion. Patients were asked to fill out a questionnaire regarding knowledge and behaviors with regard to WNV and regarding recent symptoms. Data regarding immunosuppression, blood transfusions and other transplant-related demographics were obtained electronically through the electronic transplant patient chart. Informed consent was obtained from all patients.

Laboratory testing was performed through the National Microbiology Laboratory, Health Canada and the Wadsworth Center of the New York State Department of Health. Serology was performed using HI and MIA assays, which test for total antibody (IgG and IgM) (14,15). HI reactive sera were subsequently tested by a West Nile virus IgM ELISA (16). Specificity of antibody for WNV was confirmed by plaque reduction neutralization assay (17). A random sample from 50 seronegative patients as well as all samples with West Nile virus specific antibody were tested by reverse transcriptase polymerase chain reaction (RT-PCR) using a commercially available kit (RealArt WNV RT-PCR Kit; Artus Biotech USA Inc, San Francisco, USA) as per manufacturers instructions with appropriate internal and external controls.

Descriptive summary statistics were obtained using SAS version 8.0 (Cary, NC). Exact tests and confidence intervals were requested using the FREQ procedure to test for associations and proportional differences among responses. Statistically significant differences were reported using a 5% level of significance and two-sided p-values.

Results

Patient population

A total of 833 transplant patients out of the 2360 total transplant population living in the defined geographical area were approached for consent. From this group, 12 refused consent, and blood draw was unsuccessful in five patients due to difficult venous access. Samples were, therefore, available for 816 organ transplant patients. This represented 35% of all transplant recipients living in the defined geographical area and currently being followed by the transplant program (total population = 2360 patients). Of these, 731 filled out a complete or partial survey. Baseline demographic characteristics are outlined in Table 1. Type of transplant included kidney (n = 411), liver (n = 190), lung (n = 86), heart (n = 80), pancreas (n = 43) and others (n = 6). Median time from transplant was 50.0 months (range 2 weeks–354.6 months). Maintenance immunosuppression was variable, but generally included cyclosporin or tacrolimus-based therapy in the majority of patients. In order to determine whether the sampled patients were representative of our whole population, we compared the following variables: age, gender, type of transplant, time from transplant, and immunosuppression therapy in those who were sampled versus those who were not. All variables were analyzed as categorical variables except age, and time from transplant (continuous variables). No statistically significant differences were found. Specifically, 16.9% and 18.0% on cyclosporin/prednisone in sampled versus nonsampled patients, respectively; 4.0% and 5.0% were on tacrolimus/prednisone in sampled versus nonsampled patients, respectively; 58% and 54% were on either mycophenolate mofetil or azathioprine in sampled versus nonsampled patients, respectively. Inter-quartile ranges for time from transplant for nonsampled patients were 25th percentile 20.2 months, 50th percentile 52.4 months and 75th percentile 106.2 months.

Table 1. Baseline characteristics of transplant patients
Variable Number of patients, n (%)
Age: median (range) years 51.4; range 18–86 years
Sex Male: 500/816 (61.3)
Female: 316/816 (38.7)
Type of transplant
 Kidney 411/816(50.4)
 Liver 190/816 (23.3)
 Lung 86/816 (10.5)
 Heart 80/816 (9.8)
 Kidney-pancreas 43/816 (5.3)
 Other 6/816 (0.1)
Maintenance immunosuppression (%)
 Cyclosporin/prednisone 138/816 (16.9)
 Tacrolimus/prednisone 32/816 (3.9)
 Cyclosporin/prednisone + MMF 163/816 (20.0)
 Tacrolimus/prednisone + MMF 93/816 (11.4)
 Other 390/816 (47.8)
Time from transplant (months)
 Mean ± S.E 70.62 ± 2.36
 Median (range) 50.0 (0.43 – 354.56)
 Inter-quartile ranges
  25th percentile 19 months
  50th percentile 50 months
  75th percentile 103 months

Seroprevalence

IgM antibodies specific for West Nile virus were detected in 2/816 people (seroprevalence 0.25%; 95% CI 0.03–0.88%). Both positive patients recalled a history of febrile illness in the previous 2 months for which no specific etiology was found and WNV testing had not been done. Both patients had recovered spontaneously and had attended clinic only for routine post-transplant care. Both patients reported significant outdoor recreational activity, a risk factor for community exposure to West Nile infection. One patient was a 60-year-old liver recipient 3.7 years post-transplant. The second patient was a 45-year-old liver transplant recipient 4 months post-transplant. A third patient had IgG antibodies detected against West Nile virus but no IgM. This patient was a 68-year-old female 12 years post-renal transplant who was born and had spent most of her life in Asia. This patient had no recent symptoms, and no exposure risks. Antibody against the West Nile virus nonstructural protein-5 was not detected supporting the likelihood that this individual acquired a West Nile infection prior to the 2002 outbreak (18). Therefore, this patient was suspected of having remote West Nile virus exposure rather than newly acquired infection. An additional 98 patients were positive for flavivirus antibody by HI, but plaque reduction neutralization testing was negative for West Nile virus specific antibody. A random sample of 50 patients as well as all seropositive patient samples was tested for nucleic acid by RT-PCR. All of these samples tested negative for West Nile virus nucleic acid.

Estimation of risk of neurologic disease

From the seroprevalence results and the data from hospital-based surveillance in our area that had identified four transplant patients with WNV meningoencephalitis, the risk of neurological disease in a transplant patient who acquires WNV infection could be estimated. These four patients have been described previously (13) and they all resided in the ‘at-risk’ geographical area, all developed community acquired WNV meningoencephalitis, and included liver (n = 1), kidney (n = 2) and heart (n = 1) transplant recipients. Based on a seroprevalence of 0.25% (95% CI 0.03–0.88%), and a total transplant population of 2360 patients in the defined geographical region, the predicted total number of infections in the transplant population is approximately six with a 95% CI of between 1 and 21 infections. Since hospital-based surveillance identified four cases of WNV meningoencephalitis in our transplant population, the risk of meningoencephalitis in a transplant patient with WNV infection can then be estimated as 40% (95% CI 16–80%). Similar calculations have been used in previous serosurveys in Staten Island, Queens county (New York) and Bucharest (Romania) to arrive at the meningoencephalitis risk estimate of less than 1% (approximately 1:140) for immunocompetent patients (7,8,19).

Knowledge and behaviors

Knowledge and behaviors with regards to West Nile virus were assessed by questionnaire and results are outlined in Table 2. A questionnaire was completed by 731/816 (89.6%) patients. The majority of transplant patients had previously heard of West Nile virus (89.5%). However, even though this is a population with significant exposure to the health care system and the transplant program had implemented specific educational measures, only 56% of patients knew of at least one protective measure to take against West Nile virus infection. Only 47% of patients reported using at least one protective measure and only 33% reported using an insect repellent with DEET (N,N-diethyl-3-methylbenzamide) either sometimes or often when outdoors. Although there was very little opportunity for occupational exposure to West Nile virus (only 1.7% reported working outdoors), almost half of the transplant patients reported spending a mean of one or more hours per day being outdoors at dusk or dawn, a time when mosquitoes are most active.

Table 2. Knowledge and behavior patterns in transplant recipients
Variable Number of patients (%)
Had heard of West Nile virus Yes 650/726 (89.5)
No 76/726 (10.5)
Knew at least one protective measure Yes 408/730 (55.9)
No 322/730 (44.1)
Used at least one protective measure Yes 345/731 (47.2)
No 386/731 (52.8)
Used insect repellent when outdoors Sometimes or often
 236/719 (32.8)
Never 483/719 (67.2)
Occupation
 Not working 328/587 (55.9)
 Indoors primarily 249/587 (42.4)
 Outdoors primarily 10/587 (1.7)
Length of time outdoors
 <1 h/day at dusk and dawn 376/704 (53.4)
 1–2 h/day at dusk and dawn 189/704 (26.9)
 ≥2 h/day at dusk and dawn 139/704 (19.7)
Reported seeing a dead bird Yes 127/721 (17.6)
No 594/721 (82.4)
  • Questionnaires were completed by 731/816 (89.6%) patients and missing values were present for some questions. Denominators represent those that completely answered the question.

Patients who had heard of West Nile virus were more likely to use at least one protective measure against West Nile virus (OR 5.1; 95% CI 2.7–10.2). Similarly, patients who reported knowledge of at least one protective measure were 6.6 times more likely to act on at least one protective measure (95% CI 4.7–9.4). However, patients with outdoor occupation, those who spent a mean of >2 h/day outside, or those who had reported seeing a dead bird were not any more likely to have acted on any protective measures against West Nile virus compared to those without these risk factors (p = NS for all comparisons) (Table 3).

Table 3. Factors associated with acting on at least one preventive measure
Variable Acted on at least one
protective measure
Did not act on any
protective measures
Odds ratio 95% C.I. p value
n (%) n (%)
Had heard of West Nile virus
 Yes 332 (51.1) 318 (48.9) 5.1 (2.7, 10.2) <.0001
 No 13 (17.1) 63 (82.9)
Knowledge of at least one protective measure
 Yes 271 (66.4) 137 (33.6) 6.6 (4.7, 9.4) <.0001
 No 74 (23.0) 248 (77.0)
Occupation at time of enrolment
 Primarily outdoors 8 (80.0) 2 (20.0) 4.0 (0.8, 38.8) 0.1066
 Not outdoors 289 (50.1) 288 (49.9)
Average length of time outside per day
 >2 hours 58 (41.7) 81 (58.3) 0.7 (0.5, 1.1) 0.1294
 <2 hours 278 (49.2) 287 (50.8)
Reported seeing a dead bird
 Yes 66 (52.0) 61 (48.0) 1.3 (0.8, 1.9) 0.2412
 No 274 (46.1) 320 (53.9)

Discussion

This represents the first reported seroprevalence study of West Nile virus in organ transplant recipients. The outbreak of West Nile virus in the Toronto, Canada region in 2002 provided a unique opportunity for assessment of this disease in the transplant population. Since this was the first time human West Nile virus infection was documented in Canada, the majority of our study samples were likely immunologically naïve to West Nile virus. The estimated seroprevalence in the transplant population was 0.25% (2/816) (95% CI 0.03–0.88%). Based on this data, and the number of cases of meningoencephalitis identified through hospital-based surveillance, the risk of meningoencephalitis in a transplant patient with community-acquired WNV infection can be estimated at 40% (95% CI 16–80%). This is in contrast to immunocompetent persons, in whom similar calculations have estimated the risk of severe neurological disease to be less than 1% (7,8,19). Therefore, transplant recipients appear to be at much higher risk of severe neurological disease due to West Nile virus compared to immunocompetent individuals. This represents the first scientific estimate of the risk of encephalitis in a transplant patient with community acquired WNV infection.

Other interesting findings included data on the knowledge and behavior patterns of transplant patients. This is a patient population that has significant exposure to the medical community and receives specific teaching about infectious risks as part of their routine transplant care. In addition, at the onset of the outbreak of West Nile virus, our transplant patients were delivered specific information about West Nile via an automated messaging service. This included detailed instructions about the need for personal protective measures. Despite this, only 56% of patients knew of at least one protective measure and only 47% had acted on at least one protective measure. These numbers were lower than reported for the general population in a New York-based serosurvey (7). The compliance with insect repellant use with outdoor activity was poor at only 33%, a number comparable to that reported for the general population (7). Also, it was clear, that most transplant patients are well enough and sufficiently functional to have significant opportunity for community exposure to West Nile virus. Almost half of patients reported spending 1 h or more outdoors between dusk and dawn, a time when mosquitoes are most likely to feed. Finally, those who had specific risk factors such as outdoor occupation or spending >2 h outdoors, were no more likely to use personal protection measures compared to those without these risk factors.

Seroprevalence studies for West Nile virus have been conducted in immunocompetent persons in the community and hospital settings. There are no published studies in the nontransplant setting for our specific region and thus, we can only compare our results to post-outbreak seroprevalence studies in other areas. In these studies, seroprevalence rates of less than 1% to 4.1% have been reported and have been helpful in estimating the rate of meningoencephalitis per total West Nile virus infections (7,8,19). The proportion of one meningoencephalitis case per 140 total West Nile virus infections estimated in Queens County is similar to that observed in the Bucharest, Romania epidemic and in Staten Island, New York. Age has been the only factor previously shown to be associated with an increase in the risk of meningoencephalitis. In persons greater than 65, the proportion has been estimated at 1/50, and for those less than 65 at 1/300.

Limited epidemiological data are available regarding community acquired West Nile virus infection in immunosuppressed transplant patients. West Nile infections in transplant recipients may also result from contaminated blood products (10). In a description of 23 cases of transfusion transmitted West Nile virus infection in the United States in 2002, two (8.7%) were immunocompromised due to organ transplantation. One of the patients who acquired transfusion transmitted West Nile virus infection subsequently died, and donated organs to four transplant recipients, three of whom developed meningoencephalitis (9). Although, these and other case reports have suggested that organ transplant patients may have more severe disease compared to immunocompetent patients, it is only with seroprevalence data that any true estimates regarding rates of asymptomatic and symptomatic disease can be made in this population (9,10,20,21).

We used a sampling method similar to the Bucharest, Romania serosurvey. However, unlike other studies, we sampled a large proportion of our population (35%) and we were able to compare sampled patients with nonsampled patients in terms of demographics and immunosuppressive therapy. This suggested that our sample was reasonably representative of our transplant population. A theoretical limitation of our study is that the most heavily immunosuppressed transplant recipients may not mount a serological response leading to false negative results. This may have resulted in an underestimation of the true prevalence of West Nile infection in this population. Indeed, the use of serological testing for acute diagnosis of viral infection in transplant patients is often unreliable. As such, there will remain some doubt as to the true prevalence of WNV in the population we tested. However, lack of an appropriate immune response in the face of excess immunosuppression usually predisposes to more severe presentations. This may be especially true for WNV, in which several animal studies suggest a critical role for humoral immunity in controlling infection (22–25). Together, these data suggest that antibody mediated immunity may be more important than cellular immunity. For example, in a study evaluating cyclophosphamide-treated mice challenged with WNV, immune serum was more protective than splenocytes (22). Also, a B-cell knockout mouse model is exquisitely susceptible to West Nile virus and survival in animal models can be drastically improved with specific immune globulin therapy (23–25). Given the importance of humoral immunity, we think it is unlikely that there may have been many asymptomatic transplant patients with West Nile infection but with no seroconversion. Therefore, our estimation of the rate of neurological disease in infected transplant patients should still be accurate.

Other potential limitations of our study include the nature of this cross-sectional survey, which is limited to one period in time and, therefore, may have missed some patients who acquired WNV at a later point. Also, the accuracy and reliability of the assays we used to test for antibody have not been previously tested in transplant patients. An assessment of seroprevalence data in the nontransplant population in the same geographic region would have provided a useful comparator. Finally, it is possible that the patients we sampled were not representative of our transplant population due to unknown or unmeasured confounders.

In summary, transplant recipients have a much greater risk of meningoencephalitis when infected with WNV than immunocompetent persons. This study also highlights difficulties in public health education of these patients. Despite extensive exposure to the health care system, and specific attempts at education regarding West Nile virus, knowledge of and use of personal protection measures was poor. Such measures are critical in this patient population, and transplant programs should devise strategies to ensure greater compliance to personal protection measures.

Acknowledgments

We are grateful to Dr David Grant and Dr Victoria Edge for review of the manuscript. We thank Ms Rebekah Boyle and Ms Valerie Demarest for technical assistance in laboratory investigations. We also thank Ms Jane Hamel for assistance with patient enrolment, and Mr Rob Smith for assistance with the transplant database.

Funding sources: This work was funded by the Canadian Institutes of Health Research (CIHR grant #64457) and in part by the National Institute of Allergy and Infectious Disease and National Institutes of Health, under contract N01-AI-25490.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.