Sequential virus monitoring of pediatric patients with hematopoietic stem cell transplantation by multiplex PCR method

Aim: To investigate the risk factors of virus infection and clinical significance of virus monitoring in pediatric patients with allogeneic stem cell transplantations (SCT). Methods: Peripheral blood was weekly checked for HSV-1, 2, VZV, EBV, CMV, HHV-6,-7,-8, BKV, JCV and PBV19 by multiplex PCR method until day100120 in 33 pediatric patients with 35 allogeneic SCT performed between 2004 and 2009. Occasionally, ADV was examined in cases with suspected clinical symptoms. Transplantation-related mortality (TRM) was compared with that of 35 SCT between 1998 and 2004 when virus monitoring was not introduced. Result: EBV was positive for 17cases (48.6%), CMV for 10 (28.6%) and HHV-6 for 13 (37.1%). Mean time of onset was 30th day for HHV-6, 41st day for CMV, and 46th day for EBV. HSV-1 was positive for three cases, BKV and JCV for two cases, and HSV-2, VZV, HHV-7 and ADV for one case. No virus was detected in 12 cases (34.3%). Multivariate analysis showed that ATG and HHV-6 infection were statistically relevant (odds ratio 40.03, p=0.04). GvHD III-IV was related with CMV (odds ratio 6.25, p=0.09) and EBV (odds ratio 10.47, p=0.08) infection, and reduced-intensity conditioning (RIC) was related with HHV-6 infection (odds ratio 8.65, p=0.16), although not statistically significant. To detect CMV infection, PCR was more sensitive than CMV pp65 antigenemia. Positive PCR result for CMV, EBV, or HHV-6 before SCT seemed to be a risk factor for reactivation of each virus. Positive antibody for CMV of the recipient was also a risk factor of CMV infection. TRM in 2004-2009 cohort was 0.208 and TRM in 1998-2004 cohort was 0.301, and the introduction of virus monitoring improved TRM, although not statistically significant. Conclusion: Multiplex virus monitoring before and after SCT is possibly useful for the management of pediatric patients with SCT. Further examination is necessary to confirm its clinical significance. Correspondence to: Masayuki Nagasawa, Department of Pediatrics, Musashino Red Cross Hospital 1-26-1, Kyonan-cho, Musashino-city, Tokyo 180-8610, Japan; Tel +81-422-32-3111; Fax +81-422-32-3525; E-mail: mnagasawa.ped@tmd.ac.jp


Introduction
Viral infection is still a main factor of the morbidity and mortality after the hematopoietic stem cell transplantation (SCT) while supportive care has been strengthened [1]. Furthermore, with the expansion of the transplant feasibility, the importance of opportunistic infection increases along with the development of novel immunosuppressant and conditioning regimen [2]. Prophylactic treatment reduced the incidence of early viral infection, but increased the late viral infection by compromising the anti-viral immune recovery, resulting in no improvement of overall survival [3,4]. Recently, it has been reported that early detection of targeted viruses and preemptive treatment are important to have good outcomes [5]. There are various reports in the adult SCT, but there are few examinations about the infection after pediatric SCT. From the results of viral monitoring in our institute, we investigate risk factors of virus infection after the pediatric SCT and examined the clinical impact of sequential virus monitoring by comparing with the historical control in which virus monitoring was not introduced.  (Table 1), we performed a sequential virus monitoring using the multiplex PCR method with a peripheral blood specimen once a week until 100-120 days after transplant. In 31 cases, PCR analysis was performed since one month before SCT. We analyzed urine, stool, intestinal mucosa specimen when suspected symptoms were observed.

Methods
As a historical control, 35 patients who were transplanted between 1998 and 2004 (Table 2), when virus monitoring was not introduced at our institute were analyzed.
In qualitatively positive cases, real time PCR for each virus was performed, respectively. We defined that virus was positive when virus was detected in real time PCR method as described before [6][7][8]. PCR assay was performed at Cell Therapy Center of Tokyo Medical and Dental University Hospital. Primers used in multiplex PCR and quantitative nested PCR for virus detection are listed in Table 3.
Procedure of hematopoietic stem cell transplantation and supportive care: In principle, GvHD prophylaxis consisted of cyclosporine (CSP) plus short-term methotrexate (MTX) for human leukocyte antigen (HLA)-identical-related donors, and tacrolimus (TAC) plus short-term MTX for HLA-identical unrelated and HLAmismatched-related donors. For unrelated umbilical cord blood donors, CSP or TAC and short-term MTX were used. In some cases, methylprednisolone (mPSL) or prednisolone (PSL) was added as GvHD prophylaxis. CSP was administered twice daily, and blood trough level was adjusted between 150-250 ng/mL, and TAC was continuously administered and blood level was maintained between 8-12 ng/mL. As a prophylaxis for infections, aciclovir and anti-fungal drug (fluconazole or micafungin sodium) were started at 7 days and one day before SCT, respectively. Trimethoprim-sulfamethoxazole (ST combination) was discontinued one day before SCT, and restarted after engraftment. In a case with previous Pneumocystis jirovecii infection, ST combination was continued during SCT. Immunoglobulin was administered to keep serum IgG level more than 500 mg/dL every 2 weeks. We preemptively started ganciclovir or foscarnet sodium hydrate for clinically suspected CMV infection case or case with viral load of 1 × 10 2 copies/μgDNA and more with positive CMV pp65 antigenemia (more than 5 positive cells out of 50,000 cells). Engraftment was determined as the first day when neutrophil count exceeded 500/μl for three consecutive days.

Statistics:
We calculated controlled odds ratio by the Chi-squared test and logistic regression assay (95% confidence interval). A P-value of less than 0.05 was considered statistically significant difference.
GvHD prophylaxis included CSP of one case, CSP + PSL of two cases, CSP + sMTX of 10 cases, CSP + sMTX + mPSL of one case, TAC+sMTX of 19 cases and TAC+sMTX + mPSL of two cases.

Transplant results
In 33 patients, eight patients were dead due to transplantationrelated complications. One ALL patient relapsed, received 2 nd allogeneic  SCT, and died because of severe hepatic VOD. Two patients rejected first SCT, and one successfully accepted 2 nd SCT, and the other was dead due to transplantation-related complication ( Table 1). Figure 1): In 23 cases (65.7%), one or more viruses were detected, and their average age was 9.4 years (from 7 months to 21 years; median of 11 years). In 12 cases, no virus was detected, and their average age was 3.1 years (from 8 months to 17 years; median of 1.5 years). HHV-6 was positive in 36 specimens (10.7%), EBV in 34 specimens (9.9%), and CMV in 33 specimens (9.5%). The cumulative positive incidences were EBV of 48.6%, HHV-6 of 37.1% and CMV of 28.6% (Figure 1). In the average, EBV was detected at 46.4 days, CMV at 41.1 days, and HHV-6 at 30.5 days after transplantation. In some cases, multiple viruses were detected in the same specimen at once. HHV-8 and PVB19 were not detected in any samples. BKV and JCV were positive in two cases (pt.nos.10 and 12), respectively. ADV was detected in one case transiently (pt.nos.20).

Virus
Sequence for primers and probes     CMV load of 1 × 10 2 copies/μg DNA or more, CMV pp65 antigenemia was positive (more than 5 positive cells out of 50,000 cells) in four cases who received anti-CMV agents ( Figure 2). Whereas asymptomatic three cases (pt.nos.7, 13 and 34) with CMV pp65 antigenemia negative (less than 5 positive cells out of 50,000 cells) turned negative for CMV without anti-CMV agent even if CMV load was transiently 1 × 10 2 copies/μg DNA or more after the continued inspection. Three cases (pt.nos. 6, 10 and 12) with CMV load of 1 × 10 2 copies/μg DNA or less received anti-CMV agents because of suspected symptoms.
Other than CMV, viral load became decreased or negative along with the recovery of immune function without any intervention.
EBV load was less than 1 × 10 4 copies/μg DNA in all cases, and no one developed lymphoproliferative disease.
One case (pt.nos. 8) was transplanted at 11-month-old and HHV-6 was detected as a primary infection at 96 th day after transplantation. In other four (pt.nos. 11, 19, 22 and 34: aged between 5 and 6 years), the peak of HHV-6 load was between 13 and 19 days after transplantation as a reactivation. In the SCT of adult patients, significance of HHV-6 infection is being emphasized, especially as a causative pathogen of encephalitis. Clinical significance of HHV-6 infection in the pediatric SCT is still controversial [10,11]. (Table 5): We examined 31 cases in which viral PCR analysis was performed within one month before transplantation to find out the relation with virus reactivation after SCT.

Association with virus infection status before the transplant
CMV was detected in one out of three CMV positive for PCR (33.3%) and in 8 out of 28 cases (28.6%) in which CMV was negative for PCR before transplantation. In 14 cases with positive anti-CMV antibody, CMV was detected in 8 cases (57.1%) after transplantation. Out of 11 cases in which anti-CMV antibody was negative before transplantation, CMV was detected in only one case (9.1%).
EBV was detected in 7 out of 9 cases (77.8%) in which EBV was positive for PCR before transplantation, and HHV-6 was detected in 3 out of 4 cases (75%) in which HHV-6 was positive for PCR before transplantation. On the contrary, EBV was detected in 8 out of 22 cases (36.4%) in which EBV was negative for PCR before transplantation. HHV-6 was detected in 8 out of 27 cases (29.6%) in which HHV-6 was negative for PCR before transplantation. (Table 6): To find out the risk factors for viral infection after SCT, we performed multivariate analysis. Only ATG administration and HHV-6 infection (odds ratio 40.03, p=0.04) showed a statistically significant relevance. GvHD III-IV was related with CMV (odds ratio 6.25, p=0.09) and EBV infection (odds ratio 10.47, p=0.08), and RIC was related with HHV-6 infection (odds ratio 8.65, p=0.16), although not statistically significant (Table  4). TMA was associated with CMV infection (odds ratio 7.91, p=0.09), but not with EBV (odds ratio 0.46, p=0.52) or HHV-6 infection (odds ratio 1.36, p=0.74), which is different from the report with adult SCT [12]. Unrelated cord blood was not related with HHV-6 infection.

Clinical impact of virus monitoring on TRM in pediatric SCT:
To investigate the clinical impact of virus monitoring on the outcome of SCT, we compared TRM in 2004-2009 cohort with TRM in 1998-2004 cohort. The patient profiles of 1998-2004 cohort were presented precisely in Table 2. SCT procedure and supportive care were the same as described above in Methods. Engraftment syndrome was not recorded in this cohort because the concept and criteria of ES were not accepted in this period. The content of background diseases and conditioning regimens were comparable between two cohorts. As Multivariate analysis showed that GVHD(≧3) was the most significant factor for CMV (odds ratio = 10.47) and EBV (odds ratio = 6.25) and ATG for HHV-6 (odds ratio = 40.03). *:Patients with antibody deficiency were excluded. Positive EBV or anti-CMV antibody before SCT predicts EBV or CMV infection after SCT with an odds ration of 6.13 (p=0.036) and 13.3 (p=0.013) respectively.

Discussion
It has been reported that viruses were detected from the blood in 48 to 63% of the pediatric patients until 100 days after HSCT [13,14]. HHV-6 has been reported to be detected two to four weeks earlier than CMV, and the reported frequency has varied from several to 70% [13,[15][16][17][18][19][20]. The frequency to detect CMV and EBV has been reported to be 11-28% and 24-48%, respectively [13,21,22]. In our observation, the frequency of virus detection was almost the same as previously reported. However, it was significantly lower in infants in which primary immunodeficient patients were dominantly involved. The average age of virus-positive group was 9.4 years (median of 11 years), and that of virus-negative group was 3.1 years (median 0.5 years). Especially, in the 15 patients of one year old or younger (infants group), EBV was detected in only 4 cases (26.7%), CMV in one (6.7%), and HHV-6 in three (20.0%). No virus was detected in 9 patients (60.0%). In the infants group, majority of patients had not encountered viral infections before SCT. Most of the immunnodeficient patients had been managed to be protected from infection since they were diagnosed. These patients are at risk for primary infection from the transfusions. It has been reported that CMV negative donor is preferable to avoid CMV infection in these patients [23]. Furthermore, transfusion-associated CMV is reported to be prevented by using the leukocyte-filtered blood products [24]. In our institute, CMV negative donor could not be supplied in any of the cases, but all of blood products were leukocyte-filtered. In the older children, they are already infected and reactivation of virus often becomes a clinical problem as in the adult case under the secondary immunodeficient condition after SCT [20].
We re-evaluated the correlation of CMV viral load and CMV pp65 antigenemia level, and its association with the prognosis in the pediatric SCT [25][26][27].
Three cases (pt.nos.6, 10 and 12) with CMV load of 1 × 10 2 copies/ μg DNA or less received anti-CMV agents because of suspected symptoms, although their CMV pp65 antigenemia were negative. Three in 7 cases with CMV load of 1 × 10 2 copies/μg DNA and more were CMV pp65 antigenemia negative. All of CMV pp65 antigenemia positive patients showed CMV load of 1 × 10 2 copies/μg DNA and more. As previously reported, sensitivity of CMV pp65 antigenemia is unstable and unreliable when leukocyte number is low during the early phase of SCT [26]. These results confirm that PCR testing is more sensitive than antigenemia and is more suitable for monitoring after SCT.
To start preemptive intervention, the level of 1 × 10 2 copies/μg DNA and more or CMV pp65 antigenemia of 5 and more out of 50,000 cells seemed to be reasonable practically in our experience as suggested before [27]. However, preemptive therapy should be started for the patients with suspected symptoms, even if viral DNA load is lower than cutoff level.
It has been reported that more than 1 × 10 2.5 copies/μg DNA of EBV load is a sign of viral reactivation, and more than 1 × 10 4 copies/μg DNA is a risk factor for EBV lymphoproliferative disease (EBV-LPD) [28]. In this study, four cases showed 1 × 10 2.5 copies/μg DNA or more of EBV load, but not exceed 1 × 10 4 copies/μg DNA and all of them resolved spontaneously. Clinical point of view, EBV load of 1 × 10 4 copies/μg DNA or more seems reasonable as a cutoff for preemptive intervention of EBV-LPD such as anti-CD20 antibody administration.
It has been reported that the infection history of both of donor and recipient is associated with viral infection and reactivation after SCT [13,17].
In the PCR positive case of EBV and CMV before transplant, the same virus was more likely to be detected after transplant (Table 5). For CMV infection, the antibody test of recipients before transplant was also important. Positivity of antibody for CMV was a risk of reactivation even if PCR testing before transplantation was negative. In some cases (pt.nos. 11, 22 and 30), CMV was positive in stool and pharynx, but negative in blood before SCT, and then CMV was detected in the blood after SCT. In cases of antibody-production deficiency, direct detection of virus before SCT is an only method to find out the previous or insidious infection of CMV. In the patients with immunodeficiency, viruses are often detected persistently even without clinical symptoms. The evaluation of antibody and virus before transplant is also useful and important as well as early preemptive detection of virus by the regular monitoring after SCT.
UR-BM, HLA mismatch, URCB, MAC, T-cell-depleted transplant, ATG, steroid, and GvHD have been reported to be risk factors for virus infection and reactivation [13][14][15][16]18,25]. In our study, HHV-6 infection was not related with unrelated cord blood contrary to the report from adults SCT [18]. This may be due to the fact that most of unrelated cord blood was transplanted to the patients under two years old, who might not be infected with HHV-6 yet. A statistically significant relevance was found between ATG administration and HHV-6 infection. However, CMV and EBV infection were not related with ATG. CMV and EBV infection were related with acute GvHDIII-IV, although not statistically significant. This may be interpreted that short-term intensive immunosuppression is related with HHV-6 infection/reactivation, while long-term immunosuppression combined with inflammation is prerequisite for CMV and EBV infection/reactivation. Finally, in order to investigate the clinical impact of sequential virus monitoring, we compared TRM of this cohort with that of historical control in which virus monitoring was not introduced.  patients than 1988-2004 cohort. However, we have to be cautious in the interpretation of this result because the number of patients in both cohorts was small and the profiles of both patients and conditioning regimen were not the same. Furthermore, TRM is dependent on not only virus infection but also other SCT-related risk factors such as donor source, conditioning regimen, HLA-disparity, treatment history, and so on.
Although, the introduction of virus monitoring might not have statistically significant impact of the improvement of TRM, future development of novel anti-virus therapy such as virus-specific adoptive cell therapy and anti-virus medicine will improve the clinical consequences ultimately in combination with early detection of virus infection by sequential virus monitoring. Actually, it has been reported that pre-empiric anti-CD20 administration after early detection of EBV is useful for the prevention of development of EBV-LPD after SCT [29,30].
Our observation seems instructive and further examination is necessary to resolve clinical problems with viral infection in pediatric SCT.