Accepted Manuscript Rotavirus disease course among immunocompromised patients; 5-year observations from a tertiary care medical centre P. Bruijning-Verhagen, M.D. Nipshagen, H. de Graaf, M.J.M. Bonten PII:
S0163-4453(17)30263-3
DOI:
10.1016/j.jinf.2017.08.006
Reference:
YJINF 3967
To appear in:
Journal of Infection
Received Date: 13 April 2016 Revised Date:
19 June 2017
Accepted Date: 9 August 2017
Please cite this article as: Bruijning-Verhagen P, Nipshagen MD, de Graaf H, Bonten MJM, Rotavirus disease course among immunocompromised patients; 5-year observations from a tertiary care medical centre, Journal of Infection (2017), doi: 10.1016/j.jinf.2017.08.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Rotavirus disease course among immunocompromised patients; 5-year observations from a tertiary care medical centre P. Bruijning-Verhagen a, M.D. Nipshagen b* , H. de Graaf c*, M.J.M. Bonten a,b
b
Department of Medical Microbiology, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands
c
M AN U
NIHR Clinical Research Facility Southampton, University of Southampton, C Level, West Wing, Mailpoint 218, University Hospital Southampton NHS Foundation Trust, Tremona Road, Southampton, SO16 6YD. England
SC
Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands
RI PT
a
TE D
Running title: Rotavirus among immunocompromised patients
AC C
EP
Corresponding Author: Patricia Bruijning-Verhagen Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands phone: (+31) 88 756-8137 fax: (+31) (0)88 75 68099 email: [email protected]
ACCEPTED MANUSCRIPT Summary Rotavirus (RV) is highly endemic inside and outside hospital-settings. Immunocompromised children and adults are at risk of complicated rotavirus gastroenteritis (RVGE), but general rotavirus disease severity in this group remains poorly described and rotavirus testing is not routinely performed beyond infancy. We assessed rotavirus disease among
RI PT
immunocompromised hospitalized patients. Methods: Rotavirus infections at a Dutch tertiary-care centre were identified from 5-year laboratory records. Rotavirus disease course was evaluated by chart review for each
immunocompromised patient. In a matched case-control analysis, we assessed whether being
SC
immunocompromised predisposed to RVGE. Rotavirus testing practice for suspected infectious gastroenteritis in our hospital was determined over a 3-years period.
M AN U
Results: Out of 4584 RV tests performed, 294 were positive among hospitalized patients. Immunocompromised patients represented 57% (n=20) of adult, and 12% (n=32) of pediatric RVGE. A complicated disease course occurred in 81% of them and 33% required adaptations in underlying disease management. Immunocompromised adults were 7.4 times more likely to develop RVGE compared to non-immunocompromised matched hospital-controls. Rotavirus testing in adult patients with suspected infectious gastroenteritis was uncommon (12% tested).
TE D
Conclusions: In our hospital, most adults with RVGE are immunocompromised compared to a much smaller proportion in children. RVGE in immunocompromised patients is associated with significant morbidity. Routine rotavirus testing beyond infancy should be recommended
AC C
EP
for immunocompromised patients with suspected infectious gastroenteritis.
ACCEPTED MANUSCRIPT
RI PT
Key Words: Rotavirus Immunocompromised Complications Infectious gastroenteritis Hospital Nosocomial Vaccination
AC C
EP
TE D
M AN U
SC
Highlights: • The majority of hospitalized rotavirus gastroenteritis in adults occurs in immunocompromised patients • 81% of immunocompromised rotavirus patients had a complicated disease course • Rotavirus infection interfered with underlying disease management in 33% of patients • Rotavirus testing in adults with suspected infectious gastroenteritis is uncommon (12% tested) • We recommend rotavirus testing beyond infancy for immunocompromised patients
ACCEPTED MANUSCRIPT Introduction Rotavirus (RV) causes acute diarrhoea, vomiting and fever in young children. Among adults and older children RV disease is usually the result of reinfection and generally produces only mild diarrhoea or asymptomatic infection because of pre-existing immunity.(1-4) Among immunocompromised individuals however, prolonged diarrhoea, extra-intestinal
RI PT
manifestations and associations with other gastrointestinal pathology such as Graft versus Host Disease (GvHD), mucositis and colitis have been observed suggesting that RV disease in this group is much more severe.(5-7) These manifestations though have only been described in case reports and small case-series, which do not provide insight in the frequency of
SC
occurrence of such complications or the severity of RV disease in general among immunocompromised patients.
RV is easily transmitted and resistant to many aseptic solutions and, without universal infant
M AN U
RV vaccination programs in place, is highly endemic inside and outside the hospitalsetting.(8-10) As such, RV can be an important pathogen among hospital populations hosting high numbers of immunocompromised patients. In the absence of effect on clinical management, RV testing is not routinely recommended beyond infancy and RV disease among adult hospital patients is poorly quantified.
TE D
From 2006 onwards, infant RV vaccination programs have been implemented in many countries. Observational and mathematical modelling studies have shown that infant RV vaccination programs also generate an indirect effect by reducing RV infections among unvaccinated children and adults.(11-14) Assessment of RV disease outside the typical age-
EP
group of young children, and in particular among susceptible populations such as immunocompromised patients, prior to implementation of universal rotavirus vaccination can
AC C
help to better quantify potential benefits of indirect protection due to mass vaccination, i.e. herd-protection. In the Netherlands the use of RV vaccine has been nearly absent (< 1%) in past years. It therefore provided an ideal setting to explore RV disease in these vulnerable patient-groups.
The objective of our study is to describe RV gastroenteritis (GE) in terms of severity and disease characteristics among immunocompromised hospitalized patients, in both children and adults. In addition, we investigated RV testing practices among hospitalized patients with suspected infectious GE.
ACCEPTED MANUSCRIPT Methods Setting The University Medical Centre Utrecht (UMCU) is a 1042-bed tertiary medical centre in the Netherlands. The hospital hosts paediatric and adult haematology/oncology units covering different cancer diagnoses and treatment modalities as well as stem cell transplantation units.
RI PT
In addition, the hospital offers different solid organ transplant programs including renal, heart and lung transplantation. Enteropathogen testing in hospitalized patients is ordered on clinical grounds for suspected infectious diarrhoea. Typically, bacterial stool testing and C. difficile toxin are the most commonly performed tests among hospitalized patients. Most common
SC
viral diagnostics include norovirus (NV) and RV, particularly among children. Commercially available enzyme-immuno assay for RV stool testing is used (Rotaclone®). In the Netherlands, Campylobacter jejuni, C. difficile and NV are the most common pathogens
M AN U
identified among adult GE hospitalizations and nosocomial infections, whereas RV is most commonly detected among pediatric inpatients.(15;16)
RV cases; Identification and data collection
Laboratory reports for the years 2006 to 2010 were used to identify all RV stool tests
TE D
performed and of those, the number of RV positive specimens and basic patient demographics, including outpatient or hospitalized status. Subsequently, medical records of RV positive inpatients were screened for presence of immunocompromising conditions and other underlying chronic diseases. Patients were classified as immunocompromised if they
EP
were on anti-cancer treatment, had received stem-cell transplantation within the past year, had a solid organ transplant or suffered from an immunodeficiency syndrome at the time of RV
AC C
infection or were HIV infected with CD4 cells/mm3 48 hours after
RI PT
hospital admission or within 48hrs of previous hospital discharge.(18) The reviewers compared their findings for each case, discrepancies were discussed and resolved.
Control group
SC
To assess whether RV disease among adult inpatients was more common in the presence of immunocompromising conditions we determined the immune status in a control group of hospitalized patients without GE symptoms and matched for age, admission date and ward.
M AN U
For nosocomial RV infections, we also matched for length of stay. Eligible controls were identified and randomly selected from the hospital administrative database. Two controls were included for each RVGE case.
Statistical analysis
TE D
Among immunocompromised patients with RVGE we used summary statistics to describe frequency and duration of symptoms, treatment interventions and complications. In additional stratified analyses we compared disease characteristics between adult and paediatric patients and between those with RV as the exclusive cause of gastrointestinal symptoms and those
EP
with other concomitant conditions. We used the Mann-Whitney U-test for continuous outcomes and Chi-square or Fisher’s exact test for categorical variables, where appropriate. Conditional logistic regression was used to assess if immunocompromising conditions were
AC C
more common among patients with RVGE than among matched hospital controls.
Suspected infectious GE episodes and RV testing Diagnostic testing for RV is established practice in pediatric clinical care, but is probably far less common for suspected infectious GE in adult patients. We determined the proportion of suspected infectious GE episodes among inpatients that included a RV test. For this, we used laboratory records to identify all microbiological investigations performed for NV, RV, C. difficile and Campylobacter jejuni among inpatients as a proxy for the number of suspected infectious GE episodes. Repeated or multiple pathogen testing less than 3 weeks apart in the same patient was counted as one episode. The proportion of suspected infectious GE episodes
ACCEPTED MANUSCRIPT that included a RV test was calculated for all adult and pediatric inpatients, and separately for those on oncology or hematology/immunology (H/I) wards, as these wards accommodate immunocompromised patients most frequently. We explored seasonal trends by calculating RV testing proportions by month of the year. As PCR-based NV testing became routinely available in our hospital for patient care in 2008,
RI PT
we restricted the analysis to laboratory records from 2008-2010.
Analyses were performed using R statistical software version 2.14.2 and IBM SPSS statistics
SC
version 20. The Institutional Review Board of the UMCU approved the study.
Results
M AN U
Between 2006 and 2010, 4584 RV stool tests were performed of which 845 were from adult patients and 3739 from children (Figure 1). RV was identified in 35 (4.1%) hospitalized adult and 259 (6.9%) hospitalized paediatric patients. Nosocomial infections represented 23% (N=8) and 40% (N=103) of RV episodes among adults and children, respectively. Overall, 89% of confirmed RV infections occurred between January and April. This seasonal pattern was observed in both children and adults.
TE D
Among 35 adult patients with RVGE, 32 (92%) suffered from underlying disease. Those without underlying disease were a pregnant woman, a man with travel related diarrhoea and a healthy woman with a 5-months old baby suffering from RVGE. Among 259 paediatric RVGE patients, 155 (60%) had an underlying chronic disease.
EP
Twenty adult (57%) and 32 paediatric patients (12%) were immunocompromised according to the pre-specified definition (Table 1). Recent stem cell transplantation (N=20, 38%) and anti-
AC C
cancer treatment (N=20, 38%) were the most frequent immunocompromising conditions, followed by solid organ transplantation (N=8, 15%) and common variable immunodeficiency (CVID, N=6, 12%). Two patients with CVID and one renal transplant recipient also received anti-cancer treatment. No significant differences were observed in patient demographics among immunocompromised versus non-immunocompromised adults, whereas among children, those with immunocompromising conditions were significantly older (median age 2.6 years vs. 11 months, p
S0163-4453(17)30263-3
DOI:
10.1016/j.jinf.2017.08.006
Reference:
YJINF 3967
To appear in:
Journal of Infection
Received Date: 13 April 2016 Revised Date:
19 June 2017
Accepted Date: 9 August 2017
Please cite this article as: Bruijning-Verhagen P, Nipshagen MD, de Graaf H, Bonten MJM, Rotavirus disease course among immunocompromised patients; 5-year observations from a tertiary care medical centre, Journal of Infection (2017), doi: 10.1016/j.jinf.2017.08.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Rotavirus disease course among immunocompromised patients; 5-year observations from a tertiary care medical centre P. Bruijning-Verhagen a, M.D. Nipshagen b* , H. de Graaf c*, M.J.M. Bonten a,b
b
Department of Medical Microbiology, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands
c
M AN U
NIHR Clinical Research Facility Southampton, University of Southampton, C Level, West Wing, Mailpoint 218, University Hospital Southampton NHS Foundation Trust, Tremona Road, Southampton, SO16 6YD. England
SC
Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands
RI PT
a
TE D
Running title: Rotavirus among immunocompromised patients
AC C
EP
Corresponding Author: Patricia Bruijning-Verhagen Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht Huispostnummer STR.6.131 Postbus 85500, 3508 GA Utrecht The Netherlands phone: (+31) 88 756-8137 fax: (+31) (0)88 75 68099 email: [email protected]
ACCEPTED MANUSCRIPT Summary Rotavirus (RV) is highly endemic inside and outside hospital-settings. Immunocompromised children and adults are at risk of complicated rotavirus gastroenteritis (RVGE), but general rotavirus disease severity in this group remains poorly described and rotavirus testing is not routinely performed beyond infancy. We assessed rotavirus disease among
RI PT
immunocompromised hospitalized patients. Methods: Rotavirus infections at a Dutch tertiary-care centre were identified from 5-year laboratory records. Rotavirus disease course was evaluated by chart review for each
immunocompromised patient. In a matched case-control analysis, we assessed whether being
SC
immunocompromised predisposed to RVGE. Rotavirus testing practice for suspected infectious gastroenteritis in our hospital was determined over a 3-years period.
M AN U
Results: Out of 4584 RV tests performed, 294 were positive among hospitalized patients. Immunocompromised patients represented 57% (n=20) of adult, and 12% (n=32) of pediatric RVGE. A complicated disease course occurred in 81% of them and 33% required adaptations in underlying disease management. Immunocompromised adults were 7.4 times more likely to develop RVGE compared to non-immunocompromised matched hospital-controls. Rotavirus testing in adult patients with suspected infectious gastroenteritis was uncommon (12% tested).
TE D
Conclusions: In our hospital, most adults with RVGE are immunocompromised compared to a much smaller proportion in children. RVGE in immunocompromised patients is associated with significant morbidity. Routine rotavirus testing beyond infancy should be recommended
AC C
EP
for immunocompromised patients with suspected infectious gastroenteritis.
ACCEPTED MANUSCRIPT
RI PT
Key Words: Rotavirus Immunocompromised Complications Infectious gastroenteritis Hospital Nosocomial Vaccination
AC C
EP
TE D
M AN U
SC
Highlights: • The majority of hospitalized rotavirus gastroenteritis in adults occurs in immunocompromised patients • 81% of immunocompromised rotavirus patients had a complicated disease course • Rotavirus infection interfered with underlying disease management in 33% of patients • Rotavirus testing in adults with suspected infectious gastroenteritis is uncommon (12% tested) • We recommend rotavirus testing beyond infancy for immunocompromised patients
ACCEPTED MANUSCRIPT Introduction Rotavirus (RV) causes acute diarrhoea, vomiting and fever in young children. Among adults and older children RV disease is usually the result of reinfection and generally produces only mild diarrhoea or asymptomatic infection because of pre-existing immunity.(1-4) Among immunocompromised individuals however, prolonged diarrhoea, extra-intestinal
RI PT
manifestations and associations with other gastrointestinal pathology such as Graft versus Host Disease (GvHD), mucositis and colitis have been observed suggesting that RV disease in this group is much more severe.(5-7) These manifestations though have only been described in case reports and small case-series, which do not provide insight in the frequency of
SC
occurrence of such complications or the severity of RV disease in general among immunocompromised patients.
RV is easily transmitted and resistant to many aseptic solutions and, without universal infant
M AN U
RV vaccination programs in place, is highly endemic inside and outside the hospitalsetting.(8-10) As such, RV can be an important pathogen among hospital populations hosting high numbers of immunocompromised patients. In the absence of effect on clinical management, RV testing is not routinely recommended beyond infancy and RV disease among adult hospital patients is poorly quantified.
TE D
From 2006 onwards, infant RV vaccination programs have been implemented in many countries. Observational and mathematical modelling studies have shown that infant RV vaccination programs also generate an indirect effect by reducing RV infections among unvaccinated children and adults.(11-14) Assessment of RV disease outside the typical age-
EP
group of young children, and in particular among susceptible populations such as immunocompromised patients, prior to implementation of universal rotavirus vaccination can
AC C
help to better quantify potential benefits of indirect protection due to mass vaccination, i.e. herd-protection. In the Netherlands the use of RV vaccine has been nearly absent (< 1%) in past years. It therefore provided an ideal setting to explore RV disease in these vulnerable patient-groups.
The objective of our study is to describe RV gastroenteritis (GE) in terms of severity and disease characteristics among immunocompromised hospitalized patients, in both children and adults. In addition, we investigated RV testing practices among hospitalized patients with suspected infectious GE.
ACCEPTED MANUSCRIPT Methods Setting The University Medical Centre Utrecht (UMCU) is a 1042-bed tertiary medical centre in the Netherlands. The hospital hosts paediatric and adult haematology/oncology units covering different cancer diagnoses and treatment modalities as well as stem cell transplantation units.
RI PT
In addition, the hospital offers different solid organ transplant programs including renal, heart and lung transplantation. Enteropathogen testing in hospitalized patients is ordered on clinical grounds for suspected infectious diarrhoea. Typically, bacterial stool testing and C. difficile toxin are the most commonly performed tests among hospitalized patients. Most common
SC
viral diagnostics include norovirus (NV) and RV, particularly among children. Commercially available enzyme-immuno assay for RV stool testing is used (Rotaclone®). In the Netherlands, Campylobacter jejuni, C. difficile and NV are the most common pathogens
M AN U
identified among adult GE hospitalizations and nosocomial infections, whereas RV is most commonly detected among pediatric inpatients.(15;16)
RV cases; Identification and data collection
Laboratory reports for the years 2006 to 2010 were used to identify all RV stool tests
TE D
performed and of those, the number of RV positive specimens and basic patient demographics, including outpatient or hospitalized status. Subsequently, medical records of RV positive inpatients were screened for presence of immunocompromising conditions and other underlying chronic diseases. Patients were classified as immunocompromised if they
EP
were on anti-cancer treatment, had received stem-cell transplantation within the past year, had a solid organ transplant or suffered from an immunodeficiency syndrome at the time of RV
AC C
infection or were HIV infected with CD4 cells/mm3 48 hours after
RI PT
hospital admission or within 48hrs of previous hospital discharge.(18) The reviewers compared their findings for each case, discrepancies were discussed and resolved.
Control group
SC
To assess whether RV disease among adult inpatients was more common in the presence of immunocompromising conditions we determined the immune status in a control group of hospitalized patients without GE symptoms and matched for age, admission date and ward.
M AN U
For nosocomial RV infections, we also matched for length of stay. Eligible controls were identified and randomly selected from the hospital administrative database. Two controls were included for each RVGE case.
Statistical analysis
TE D
Among immunocompromised patients with RVGE we used summary statistics to describe frequency and duration of symptoms, treatment interventions and complications. In additional stratified analyses we compared disease characteristics between adult and paediatric patients and between those with RV as the exclusive cause of gastrointestinal symptoms and those
EP
with other concomitant conditions. We used the Mann-Whitney U-test for continuous outcomes and Chi-square or Fisher’s exact test for categorical variables, where appropriate. Conditional logistic regression was used to assess if immunocompromising conditions were
AC C
more common among patients with RVGE than among matched hospital controls.
Suspected infectious GE episodes and RV testing Diagnostic testing for RV is established practice in pediatric clinical care, but is probably far less common for suspected infectious GE in adult patients. We determined the proportion of suspected infectious GE episodes among inpatients that included a RV test. For this, we used laboratory records to identify all microbiological investigations performed for NV, RV, C. difficile and Campylobacter jejuni among inpatients as a proxy for the number of suspected infectious GE episodes. Repeated or multiple pathogen testing less than 3 weeks apart in the same patient was counted as one episode. The proportion of suspected infectious GE episodes
ACCEPTED MANUSCRIPT that included a RV test was calculated for all adult and pediatric inpatients, and separately for those on oncology or hematology/immunology (H/I) wards, as these wards accommodate immunocompromised patients most frequently. We explored seasonal trends by calculating RV testing proportions by month of the year. As PCR-based NV testing became routinely available in our hospital for patient care in 2008,
RI PT
we restricted the analysis to laboratory records from 2008-2010.
Analyses were performed using R statistical software version 2.14.2 and IBM SPSS statistics
SC
version 20. The Institutional Review Board of the UMCU approved the study.
Results
M AN U
Between 2006 and 2010, 4584 RV stool tests were performed of which 845 were from adult patients and 3739 from children (Figure 1). RV was identified in 35 (4.1%) hospitalized adult and 259 (6.9%) hospitalized paediatric patients. Nosocomial infections represented 23% (N=8) and 40% (N=103) of RV episodes among adults and children, respectively. Overall, 89% of confirmed RV infections occurred between January and April. This seasonal pattern was observed in both children and adults.
TE D
Among 35 adult patients with RVGE, 32 (92%) suffered from underlying disease. Those without underlying disease were a pregnant woman, a man with travel related diarrhoea and a healthy woman with a 5-months old baby suffering from RVGE. Among 259 paediatric RVGE patients, 155 (60%) had an underlying chronic disease.
EP
Twenty adult (57%) and 32 paediatric patients (12%) were immunocompromised according to the pre-specified definition (Table 1). Recent stem cell transplantation (N=20, 38%) and anti-
AC C
cancer treatment (N=20, 38%) were the most frequent immunocompromising conditions, followed by solid organ transplantation (N=8, 15%) and common variable immunodeficiency (CVID, N=6, 12%). Two patients with CVID and one renal transplant recipient also received anti-cancer treatment. No significant differences were observed in patient demographics among immunocompromised versus non-immunocompromised adults, whereas among children, those with immunocompromising conditions were significantly older (median age 2.6 years vs. 11 months, p
