THE VALUE OF ISOLATION PROCEDURES FOR CYTOMEGALOVIRUS INFECTIONS IN CHILDREN

WITH LEUKEMIA FREDERICK Cox, MD, A N D WALTER T. HUGHES, M D

Standard contagious isolation procedures were used for 83 patients with acute lymphocytic leukemia and serologic and/or cultural evidence of cytomegalovirus infection. The infection rate, as determined serologically, for 9 months before and 13 months during isolation procedures was not decreased. Since the techniques employed were not helpful in preventing infection with cytomegalovirus in the immunosuppressed host, they have been discontinued at this hospital.

CUWH 36~1158-1161, 1975.

OSTN;\T,\I.LY ACVUIRED CYTOMEGALOVIRUS P (ChlV) infections presumably occur after contact with infected secretions a n d after blood transfusions or organ transplantation.' T h e im- munosuppressed host is at high risk for acquir- ing serious disease from C h l V infection.' Although thc exact mode of transmission has not been established, the clustering of children with leukemia in specialized clinics and hospi- tals provides a precarious resource of highly susceptible hosts, combined with cohorts who have a high prevalence of virus excretion. If patient-to-patient transmission is important in the spread of infection, isolation of C M V ex- cretors would be warranted, provided the techniques employed proved to be a n effective means for prevention of infection in susceptible individuals. S tandard contagious isolation procedures are cumbersome, somewhat expen- sive, and impinge on the flexible use of clinic and hospital rooms. T h e present study ,was under- taken to determine the effectiveness of isolation techniques during outpatient clinic visits and

From the Infectious Diseases Service, St. Jude Children's Research Hospital. hlemphis. '1".

Supportcd by Childhood Cancer Research Center Grant C:r\-U848O and XI ultidisciplinary Cancer Research Training Grant (:A-05176, National Cancer , Institute. General Research Support (;rant RR-05584 from the Division of Research Resources. National Institutes of Health, and by ALSAC:.

Address lor reprints: Frederick Cox. hlD, St. Jude Children's Research Hospital. 332 North Lauderdale. P.O. Box 318. hlemphis, TN 38101.

l ' he authors are indebted to l l r s . Juliette Shade and Mr. (;ary Grant lor their technical assistance during this study.

Kcccivcd for publication ; \ u p s t 9, 1974.

hospitalization of children with acute lympho- cytic leukemia (ALL) a t St. Jude Children's Research Hospital.

h ; 1 ~ . I e ~ i A L s I\ND METHODS

Between January, 1972 and November, 1973, 83 children with ALL had Ch4V complement- fixation (CF) antibody titers performed at the time of the first clinic or hospital admission and subsequently at variable intervals. Patients were not included in the study if insufficient numbers of CF tests were done to determine the time of seroconversion accurately. Patients were fol- lowed throughout the study or until seroconver- sion occurred. A group of 9 children who were initially seropositive were followed separately. All seroconverters, initially seropositive in- dividuals, and a randomly selected group of seronegative patients had urine samples cultured for ChlV on one or more occasions. Seronegative patients excreting virus were ex- cluded in determinations of attack rates, since the time of acquisition of infection was un- known.

During the tirst 9 months, until October, 1972, no isolation procedures were used. After that time, patients with initial antibody titers 2 1 : 8 or who underwent seroconversion or were found to be virus excretors were treated by con- tagious isolation. During inpatient or out- patient visits, hospital personnel were required to wear a gown and mask, with handwashing before and after examination of the patient. A mask was worn by the patient during transport

1158

within the hospital and during radiographic or other special procedures. During outpatient visits, patients wore paper face masks in the hallways, were kept in separate waiting areas, received medications and were examined in specially assigned rooms, and were not allowed to dine in the main cafeteria. Outpatient or in- patient rooms were cleaned and not used for 12 to 18 hours after an infected patient was dis- charged. Discarded materials in the rooms were placed in plastic bag containers which were closed before removal from the room and were taken in a second bag directly to an incinerator. Examination instruments were assigned to the rooms. Chart covers were labeled to identify patients for isolation, and an active list of in- fected patients was distributed to physicians and nurses.

ChlV CF titers were performed by the stan- dard microtechniquee using the AD-169 strain of virus as antigen (Flow Laboratories, Rock- ville, hlD, #9-461W). A titer 2 1 : 8 was con- sidered positive; a seroconversion was dcter- mined at the time of the first positive titer in a previously seronegative individual. Serial serum samples were not tested at the same time.

Koutinely collected urine samples were cultured after the addition of 10,000 units of aqueous penicillin G, 2000 pg of streptomycin sulfate, and 20 pg of amphotericin B (E.R. Squibb, #43760) to each milliliter. This mixture was kept at 4°C for 1 hour and inoculated in 0.3- ml amounts into each of four 16 X 150 mm test tubes containing human embryonic lung (HEL) fibroblasts growing in tissue culture. Eagle's medium (GIBCO #143EG) containing aqueous potassium penicillin (50 U/ml), streptomycin sulfate (50 pg/ml), mycostatin (25 U/ml), and 2% fetal bovine serum was used for maintenance of the cells. The medium was changed after ad- sorption for 1 hour at 37OC and twice weekly thereafter. ChIV was identified by the focal cytopathic effect of large rounded cells. Isolates were passed into Leighton tubes with HEL fibroblasts, stained with hematoxylin and eosin, and identified by the characteristic intranuclear and intracytoplasmic inclusions.

All seronegative patients not excreting virus were presumed to be at risk for infection. Since contact with infected individuals in and out of the hospital is an important determinant of acquisition of infection, the number of inpa- tient/outpatient days was determined for com- parison between groups of patients. Each clinic visit or inpatient day is considered a hospital

No. 3 ISOLATION FOR CYTOMECALOVIRUS INFECTIONS Cox and Hughes 1159

contact day. The mean number of hospital con- tact days per group divided by the mean period of observation in months provides a comparison of possible exposure times.

All patients were enrolled in the Total VII I Protocol for ALL at St. Jude Children's Research Hospital. Since patients are ran- domized to four groups to receive maintenance chemotherapy consisting of one to four drugs (Group I , methotrexate [MTX]; Group 2, MTX and 6-mercaptopurine [ 6 MP]; Group 3, R.lTX, 6 MP, and cyclophosphamide; Group 4, M T X , 6 MP, cyclophosphamide, and cytosine arabinoside), comparison of the relative degree of immunosuppression could be made on this basis. The radiotherapy for each group is similar.

R ESULTS

During the 9 months without isolation techni- ques, 37 children were studied, with a mean of 1.6 CF tests done per patient at a mean fre- quency of 2.7 months; and 5 of 37 (13.5%) un- derwent seroconversion (Table 1). Before seroconversion, each patient received two to six (mean 3.8) blood transfusions and had been im- munosuppressed from 2 to 5 (mean 3.0) months. Three additional patients were initially seropositive.

In the 13 months of isolation procedures, 75 patients were studied, with an average of 3.0 CF tests per patient at a mean frequency of 2.2 months; 12 of 75 (16.0%) underwent seroconver- sion (Table I ) . Before seroconversion, each child received zero to eight (mean 3.1) transfusions and had been immunosuppressed from 2 to 11 (mean 5.7) months. Six additional children were initially seropositive during this time. The relationship of seroconversion to isolation techniques and the duration of immunosuppres- sion for each seroconverter is shown in Fig. 1. A

T,\ii[.E 1. Results of Isolation 'Techniques and C F Tests for ChIV in 80 Children with ALL

No isolation With isolation

Number of patients 37 75+ %lean no. CF tests/

hlean frequency of CF

Seroconverters 5 12 ;\!tack rate 13.5% 16%

pat ient I .6 2.9

testing (mo) 2.1 2.8

* Includes 32 susceptible patients from the group prior to isolation.

I I60 CANCER September 1975 Vol. 36

total of 25 of 83 (37.5Y0) patients were seropositive.

T h e maintenance chemotherapy for ALL was similar in the groups before and during isolation (Table 2) but a greater percentage of seroconverters (84%) received three or four drugs for maintenance than nonseroconverters (49%). The number of inpatient and outpatient hospital contact days per patient per month was also similar before and during isolation, but seroconverters had 2.9 more outpatient contact days per month than nonconverters (Table 3).

C u l t u r ~ l studies revealed that 1 1 of 25 (44%) seropositive individuals were excreting virus in the urine. This includes 4 of 9 initially seropositive children who were tested after isolation procedures had begun and had been immunosuppressed for 1 to 5 months. One

' I . A I I I . E 2. Coinparison between Maintenance (:hernotherapy for ALL and Isolation 'I'echniques

and Seroconversion for ChlV

Number of milintcniince

drugs

I 2 3 4

'I'utal number of patients

Before isolation

13 (35%) 5 (14%)

I2 ( 3 2 ' F O ) 7 (19%)

37

Number of patients Non-

During Sero- sero- isolation convertem converters

22 (29r,j 2 ( i m ) 20 (82%) 12 (I6'"o) I ( 6%) 12 (19%)

IS ( 2 0 1 ) 4 (24%) 12 (19%) 26 (35%) 10 (60%) 19 (30%)

75 17 68

3 FIG. 1. Seroconversion (CF) for cytornegalovirus before and during isolation in 80 patients with acute lymphocytic leukemia. Initial CF titers were determined at the time of diagnosis of leukemia. Each line represents one patient, and the length indicates the duration of irn- munosuppression.

seropositive patient died before cultures could be obtained. Three of 55 (5.5%) seronegative patients were tested and a total of 14 of 83 (16.9%) patients had viruria.

DISCUSSION T h e prevalence of ChIV infection as deter-

mined by serologic or cultural methods was similar to other s t u d i e ~ . ~ ~ ~ ' T h e serologic deter- mination of Ch,lV infection may not detect all actively infected individuals, since up to 1 1 % of patients with leukemia may be seronegative and excreting virus.' However, CMV excretion was present in 44% of seropositive patients and may occur in 62%' to 74%' of seropositive children with leukemia. Thus, seroconvecsion often signifies the recent acquisition of infection, although in patients with leukemia the CF titer may rise and fall to negative on repeated oc- casions.' T h e chances of missing a seroconver- sion were minimized in this study by repeated antibody determinations.

The routine methods of contagious isolation of patients used in this study were ineffective in changing the infection rate of CMV, as deter- mined serologically in children with ALL. T h e two patient groups, before and during isolation, were similar in the degree of immunosuppres- sion received and the number of hospital contact days per month, suggesting equal susceptibility and hospital exposure to infection. T h e greater percentage of seroconverters receiving combina-

No. 3 ISOLATION FOR CYTOMECALOVIRUS INFECTIONS c o x and Hughes 1161

tions of three or four drugs for maintenance chemotherapy of their ALL implies that the degree of immunosuppression is related to the acquisition of infection. T h e greater number of outpatient hospital contact days among seroconverters suggests that this type of ex- posure may be important in the spread of infec- tion. However, the small number of patients in this group compared to the nonseroconverters and lack of cornparisoil with other groups of in- fected and noninfected patients permits no defi- nite conclusions.

A previous study of inpatient cubicle isolation of leukopenic patients also did not prevent CMV infection. T h e reasons for these failures are probably related to acquisition of infection from other sources such as blood transfusions and ex- posure to other asymptomatically infected children in the hospital, community, or home. Reactivation of latent ChlV infection was also possible. as evidenced by Ch,lV excretion in 4 of 9 initially seropositive children.

TABLE 3. Comparison of Inpatient and Outpatient Hospital Contact Days between Study Groups

hlean number of hospital contact days/patient/month

Before During Sermon- Nonsero- isolation isolation verters converters

Inpatient 1.2 0.9 1.7 1.6 Outpatient 3.1 3.0 7.6 4.7

As a result of this study, isolation procedures are no longer used in this hospital for patients with ChlV infections. I t is realized, however, that studies utilizing both cultural and serologic methods of diagnosis and devices such as laminar flow rooms' are needed before the pos- sibility of effective isolation can be dismissed en- tirely. However, the use of "life island" laminar flow units for this purpose is impractical in out- patient and inpatient facilities for the numerous CMV excretors expected to be encountered.

REFERENCES

I . :\rmstrong, D.. Haghbin. hl . . Balakrishnan. S. L.. and hlurphy, h1.L. ; :\syniptomatic cytomegalovirus infection in children with leukemia. Am. .7. Dis. Chifd. 122:404, 1971.

2. Benyesh-hlelnick. h l . , Dessy, S. I . , and Fernbach. I). J . : Cytomegaloviruria in children with acute leukemia and i n other children. f'rirr. .Vor. E Y ~ . H i d .\fd 117:624. 1964.

3 . Hotiry. C;. P.. Itodriquez, I... Freireich. E. J. and Frei, E.. I l l : Protected environment. prophylactic antibiotics and cancer chemotherapy. Recent Arsufts Cancer Res. 29: 16, 1970.

4. Henson. D., Siegel, S. L).. Fucillo, L). A,, XIatthew, E. , and Levine. A. S. : Cytomegalovirus infections during childhood leukemia. J . Infect. Dis. 126:469, 1972.

5. Jameson. B.. Lynch. J.. Gamble. D. K.. and Kay, H. E. h l . : Five-year analysis of protective isolation. I.nnrr/ i:1034. 1971.

6. Swer, J. L: Applications of a microtecnnique to viral serological investigations. J . fmmunol. 88:320, 1962.

7. \Veller. '1'. H. : The cytomegaloviruses-Ubiquitous agents with protean clinical manifestations. .,V. h g / . j' . t r d . 28.5:203-211, 267-274, 1971.