Clin Investig (1992) 70:503-507

OriginN ArticUe Clinical

Investigator © Springer-Verlag 1992

Comparison of shell viral culture and serology for the diagnosis of human cytomegalovirus infection in neonates and immunocompromised subjects

B. Weber, A. Hamann, B. Ritt, H. Rabenau, W. Braun, and H.W. Doerr Abteilung ffir Medizinische Virologie, Zentrum der Hygiene, Universit/its-Kliniken Frankfurt/Main

Summary. The present retrospective study com- pares the laboratory diagnosis of cytomegalic in- clusion disease (CID) by the use of "shell vial cul- ture" [i.e., immunoperoxidase staining of human cytomegalovirus (HCMV) early antigen in human fibroblasts 24 h postinoculation] to the results of serology (i.e. immunoglobulins IgG, IgM, and IgA HCMV antibody testing) in 21 infants with con- genital or postnatally acquired HCMV infection, 5 patients with lymphoproliferative disorders, 35 human immunodeficiency virus (HIV)-seropositive patients who met the Centers for Disease Control (CDC) criteria for stages IVA and IVB of HIV infection, and 115 patients suffering from the ac- quired immunodeficiency syndrome, AIDS (stages IVC-IVE according to CDC criteria). HCMV in- fection was diagnosed by means of the shell vial culture inoculated with patient samples (e.g., urine, bronchoalveolar lavage, induced sputum, etc.) and serology in 163 (92.6%) and 65 (36.9%) patients, respectively. Viral shedding was detected by shell vial culture in 100% of the neonates, 80% of the patients suffering from lymphoproliferative dis- orders, 100% of the AIDS related complex (ARC) and 89.6% of the AIDS patients. In contrast, sero- logic testing for HCMV-specific antibodies was positive in only 28.6%, 42.9%, and 34.8% of the neonates, ARC, and AIDS patients, respectively. In lymphoma patients, serologic testing gave ident- ical results (80%) to the shell vial culture tech- nique. With the use of the shell vial procedure, active HCMV infection in immunocompromised subjects and neonates can be recognized more re- liably than by serologic testing. Nevertheless, in a low percentage of patients (7.4%), virus isolation by the shell vial culture may fail to detect HCMV infection.

Abbreviations: ARC = AIDS related complex; AIDS = acquired immunodeficiency syndrome; BAL = bronchoalveolar lavage; CDC = Centers for Disease Control, Atlanta, USA; CID = cy- tomegalic inclusion disease; ELISA=enzyme-linked immuno- sorbent assay; HCMV = human cytomegalovirus ; HIV = hu- man immunodeficiency virus; IgG, IgM, IgA = immunoglobu- lins G, M, and A; SD =standard deviation

Key words: Human cytomegalovirus - Neonates - Acquired immunodeficiency syndrome and AIDS related complex patients

Human cytomegalovirus (HCMV) infection occurs in more than one-half of the subjects suffering from the acquired immunodeficiency syndrome (AIDS) at some point during the natural course of human immunodeficiency virus (HIV) infection. HCMV is also currently recognized as the most common agent of pre- and perinatal infection [7, 13, 26]. Rapid and specific detection of active HCMV infection is important for more than one reason: (i) AIDS patients often fail to react sero- logically in an appropriate way to an acute HCMV recurrence or even primary infection [3, 25]. (ii) Antiviral agents (ganciclovir, foscarnet) are clini- cally effective for the treatment of certain forms of cytomegalic inclusion disease (CID) [1, 4, 17]. (iii) Conventional virus isolation is limited by the slow development of HCMV cytopathology in cell culture [6, 12].

The present study was undertaken to compare rapid culture with serologic testing for HCMV- specific immunoglobulin IgG, IgM, and IgA anti- bodies in order to evaluate retrospectively the rela- tive utility of these techniques for the diagnosis of an active HCMV infection in 4 groups of pa- tients: infants with congenital or postnatally ac- quired HCMV infection, lymphoma patients, HIV- seropositive patients who met the Centers for Dis- ease Control (CDC) criteria for stage IVA and IVB, and patients suffering from AIDS.

Methods

Clinical samples

The processing of clinical samples (bronchoalveo- lar lavage samples, BAL, induced sputum, biopsy specimens, spinal fluid, urine samples, and swabs) was performed as described elsewhere [25].

504

Table 1. Patient characteristics

Patient group n Mean age Length of Number of Number of in years follow-up samples tested serological tests ( ± S D ) (weeks)

Urine Sputum BAL Others IgG IgM IgA

Neonates 21 0.6_+ 0,3 7 75 - 1 9 62 63 15 Lymphoma patients 5 6.0_+ 3,3 17 33 1 3 3 47 47 7 ARC patients 35 39.0 ± 10,9 25 50 47 4 9 70 67 64 AIDS patients 115 40.1_+ 8.3 17 339 143 19 60 371 341 347

Total ]76 497 191 27 81 550 518 433

BAL = Bronchoalveolar lavage; AIDS = aquired immunodeficiency syndrome; ARC = AIDS related complex; Ig = immunoglobulin

Shell vial culture

The early detection of HCMV in cell culture with monoclonal antibodies (shell vial culture) after in- oculation with patient samples was performed ac- cording to the method of Schacherer et al. [25].

Serological assays

All serological investigations were performed with the commercially available enzyme-linked immu- nosorbent assay (ELISA) kit Enzygnost-Zytome- galie (Behringwerke, FRG). All sera were pro- cessed as previously described [8, 10].

Patients

A total of 176 patients hospitalized at the Universi- ty Clinics of Frankfurt/Main in the 3-year interval from March 1988 to March 1991 were considered for retrospective analysis (Table 1). All the patients were suffering from active HCMV infection. Ac- tive infection was defined after the criteria given by Marsano et al. [18].

Results

Clinical findings

According to the criteria defined by Marsano et al. [18], all of the 176 subjects regrouped in this study had evidence of active HCMV infection. Clinical manifestations attributable to HCMV were present in 164 (93.2%). A total of 12 (6.8%) patients (3 neonates, 5 ARC and 4 AIDS patients) presented asymptomatic active HCMV infection.

In AIDS patients, HCMV chorioretinitis (n= 38) and pneumonia (n = 37) were the predominant clinical manifestations. HCMV retinitis was exclu- sively observed in AIDS patients. HCMV-associat- ed colitis was diagnosed in 2 AIDS patients and

1 lymphoma patient. Acute fever, chronic fever, and fever of indeterminate localization not attrib- utable to another pathogen were diagnosed in 8 (4.4%), 16 (8.7%), and 24 (13.0%) patients, respec- tively.

Hepatitis was a common finding in infected in- fants (n=9). Encephalitis and polyradiculitis at- tributable to HCMV were observed in 7 and 5 AIDS patients, respectively. One AIDS patient had HCMV gastritis documented by a positive culture of gastric mucosal tissue obtained by biopsy.

Results of laboratory tests

The results of serial cultures and serologic tests of the specimens from each of the 176 subjects with active HCMV infection are shown in Tables 2 and 3. In all, 163 (92.6%) patients had a positive HCMV culture from one or more sites, including 120 of 168 (71.4%) with one or more positive urine

Table 2. Comparison of shell vial culture and serology in neo- nates and immunocompromised individuals with active HCMV infection (in brackets % of cases)

Patient group n Number of patients with positive/ significant a result(s) in:

Shell vial Serology (%) culture (%)

Neonates 21 21 (100) 6 (28.6) Lymphoma patients 5 4 (80.0) 4 (80.0) ARC patients 35 35 (100) 15 (42.9) AIDS patients 115 103 (89.6) 40 (34.8)

Total 176 163 (92.6) 65 (36.9)

a Presence of IgM antibody and/or fourfold titer rise of IgG antibody not related to passive immunization and/or fourfold titer rise of IgA antibody and/or IgA antibody titer superior to 1/640

5O5

Table 3. HCMV detection by shell vial culture technique compared with HCMV serology in neonates and immunocompromised patients (n = 176) with active HCMV infection

Patient group n Number of patients with HCMV early antigen detection Number of patients with positive or in the cell culture/patients investigated (%) significant serological tests/

patients investigated (%)

Urine Sputum BAL Others IgG a IgM b IgA c

Neonates 21 21/ 21 (100) -/ - @) 1/ 1 (100) 1/ 5 (20.0) 5/ 21 (23.8) 7/ 21 (33.3) 0/ 7 (0) Lymphoma patients 5 4/ 5 (80.0) 0/ 1 (0) 1/ 2 (50.0) 0/ 1 (0) 4/ 5 (80.0) 4/ 5 (80.0) -/ (-) ARC patients 35 17/ 31 (68.0) 20/26 (76.9) 2/ 4 (50.0) 1/ 4 (25.0) 4/ 35 (11.4) 1/ 33 (2.9) 10/ 32 (28.6) AIDS patients 115 78/111 (70.3) 46/70 (65.7) 12/18 (63.2) 7/40 (0.0) 6/115 (5.2) 15/111 (13.5) 25/113 (21.7)

Total 176 120/168 (71.4) 66/97 (68.0) 16/25 (64.0) 9/50 (18.0) 19/176 (10.8) 27/170 (15.9) 35/152 (23.0)

BAL = bronchial lavage fluid; HCMV = human cytomegalovirus a IgG seroconversion or fourfold titer raise b detectable IgM antibody c IgA fourfold titer rise or titer superior to 1/640

cultures, 66 of 97 (68.0%) with one or more spu- tum cultures, and 16 of 25 (64.0%) with one or more positive BAL cultures. Overall, HCMV infec- tion was detected by shell vial culture in 310 (38.9%) out of 796 clinical specimens investigated. In contrast, fourfold titer rises or increased Ig lev- els were detected in 133 (23.7%) of a total of 562 serum samples investigated. Consequently, only 65 (36.9%) of the investigated patients presented sero- logical signs of HCMV infection.

HCMV was isolated from at least one urine sample in each infant with congenital or postnatal- ly acquired HCMV infection. Seven of 21 infants showed increased HCMV-specific IgM levels or a four-fold titre increase for specific IgG. IgA di- rected against HCMV was not detected in this group of patients.

Shell vial culture and serology detected HCMV infection in 4 of the 5 lymphoma patients. All pa- tients but 1 showed HCMV-specific IgG titre rise and/or detectable IgM against HCMV. One serone- gative patient who was submitted to a bone mar- row transplantation showed no serological signs of an active HCMV infection despite the isolation of HCMV in one bronchial lavage and the presence of an interstitial pneumonia attributable to HCMV.

HCMV was isolated from 35 (100%) and 103 (89.6%) of the 35 ARC and 115 AIDS patients, respectively. A fourfold increase of HCMV-specif- ic IgG was detected in 4 (11.4%) ARC- and 6 (5.2%) AIDS patients. Only 1 (2.9%) ARC but 15 (13.5%) AIDS patients showed detectable levels of specific IgM. The most frequently detected sero- logic marker was IgA against HCMV, since titer rises or increased levels were detected in 10 (28.6%)

and 25 (21.7%) of the two clinical groups of HIV- l-seropositive patients suffering from recurrent HCMV infection.

Effectiveness of serology and shell vial culture testing

The effectiveness of shell vial culture and serologic testing were compared in 9 neonates suffering from hepatitis and in two clinical groups of AIDS pa- tients: 38 patients with retinitis, 37 patients with interstitial pneumonia, and 9 patients suffering from gastrointestinal manifestations including esophagitis, gastritis, gastric ulceration, duodeni-

• tis, enteritis, colitis, and proctitis. An average of 7 biopsies were performed in the involved sites from each of the 8 patients of this last clinical group.

In all 9 infants suffering from hepatitis attribut- able to neonatal HCMV infection, HCMV was iso- lated from one or more urine samples, whereas serologic testing detected HCMV infection in only one-third (n = 3) of the investigated neonates.

HCMV was isolated in the shell vial culture of urine samples from 32 (84.2%) of 38 AIDS pa- tients suffering from HCMV retinitis. Serological signs of HCMV retinitis were present in only 16 (42.1%) patients. In 37 cases of interstitial pneu- monia, HCMV was isolated in the induced sputum of 33 (89.2%) patients. In contrast, elevated IgG, M, and A levels were only detected in 14 (37.8%) patients suffering from pneumonia.

HCMV was isolated from one or more biopsies in each of the 8 patients with gastrointestinal mani- festations attributable to HCMV infection. Sero- logic testing was positive in 2 individuals (25%).

506

Discussion

The great majority of studies comparing virus iso- lation and serology for the diagnosis of HCMV infection in immunocompromised patients are somewhat limited by small sample sizes [18-20, 25]. In contrast to previous publications, our study provided extensive clinical, virologic, and serologic data obtained from different patient groups at high risk of fatal HCMV infection. Despite the facts that patients were not randomly selected and a superior number of clinical specimens were investi- gated by the shell vial culture compared with the quantity of serum samples tested by serology, the differences we observed were statistically signifi- cant. The shell vial culture technique detected HCMV infection in 92.6% of the cases, whereas serologic testing was positive in only 36.9% of all the patients investigated (Z 2 test, P<0.001). The results obtained by our study showed that the sen- sitivity of the shell vial culture performed with var- ious clinical samples is comparable to that of vire- mia detection. The detection of viral shedding in the urine showed a good diagnostic sensitivity (84.2%) for HCMV-induced chorioretinitis and represents a noninvasive and powerful technique for the measurement of the clinical response to antiviral chemotherapy with ganciclovir or foscar- net.

In the present report, a considerably larger amount of sputum samples than BAL (191 versus 27) was investigated. Sputum induction with hy- pertonic saline represents a less invasive procedure than BAL and yields specimens from which HCMV isolation by the shell vial culture [23] and Pneumocystis carinii detection [2, 21-23] can be successfully made. The sensitivity of recovering HCMV from induced sputum versus BAL fluid ranges from 60% to 68%, and the specificity ranges from 85% to 100% [23]. In our study, HCMV was detected in induced sputum samples in 89.2% of the AIDS patients suffering from in- terstitial pneumonia.

Retinitis and pneumonitis were the most com- mon clinical features attributable to HCMV infec- tion in AIDS patients. HCMV retinitis was only diagnosed in HIV-l-infected patients with full- blown AIDS. According to other authors, HCMV retinitis tends to occur in the later stages of HIV infection (25%-30% of AIDS patients), probably when the immune response is most compromised; it is occasionally the initial manifestation of AIDS [14-16].

According to previous studies [9, 11, 24] HCMV-specific IgA proved to be a superior sero-

logic marker than IgM for active HCMV infection in immunocompromised patients since increased levels of specific IgA were detected in 10 ARC and 25 AIDS patients, whereas only 1 ARC patient and 15 AIDS patients showed specific IgM anti- body.

Surprisingly, only 36.9% of the 176 subjects investigated produced a detectable serologic re- sponse to HCMV infection. Chiba et al. [5] found that up to 44% of asymptomatic, congenitally in- fected infants fail to show an IgM response. It is therefore clear that to diagnose active HCMV in- fection in neonates, shell vial culture is mandatory. In contrast, other authors [27] suggest that serolog- ic testing including the detection of IgG, M, and A has a considerably higher detection rate of HCMV infection in immunocompromised individ- uals. In conclusion, shell vial culture proved to be superior to serologic testing for the laboratory di- agnosis of cytomegalovirus infection in immuno- compromised patients and neonates. Nevertheless, in a low percentage of patients (7.4%), virus isola- tion by the shell vial culture may fail to detect HCMV infection.

Acknowledgments. We thank Prof. Dr. reed. W. Stille and Prof. Dr. reed. H.B. Helm, Department of Internal Medicine, for the clinical data concerning ARC and AIDS patients.

References

1. Balfour HH Jr (1990) Management of cytomegalovirus dis- ease with antiviral drugs. Rev Infect Dis $7 : $849-$860

2. Bigby TD, Margolskee T, Curtis JL (1986) The usefulness of induced sputum in the diagnosis of Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome. Am Rev Respir Dis 133:515-518

3. Braun W, Hehr B, Rabenau H, Doerr HW (1987) Serologi- sche Diagnostik der Zytomegalie bei immunsupprimierten Patienten. AIDS Forsch 11:634-638

4. Buhles WC, Mastre BJ Jr, Tinker A J, Strand V, Koretz SH, and the Syntex Collaborative Ganciclovir Treatment Study Group (1988) Ganciclovir treatment of life- or sight- threatening cytomegalovirus infection: experience in 314 immunocompromised patients. Rev Infect Dis $3:$495- $506

5. Chiba S, Kamada M, Hanazono H, Motokawa T, Nakao T, Komori A (1979) IgG antibody against early antigen in subclinical congenital cytomegalovirus infection. Lancet I:496

6. Dieterich DT, Rahmin M (]991) Cytomegalovirus colitis in AIDS: presentation in 44 patients and a review of the literature. J Acquir Immune Defic Syndr 4:$29-$35

7. Doerr HW (1987) Cytomegalovirus infection in pregnancy. J Med Virol 17:127-132

8. Doerr HW, Geiger S (1988) Optimierung der quantitativen Antik6rpermessung mit dem ELISA unter Berficksichtigung der Klinischen Plausibilitfit. Lab Med 12:14~146

9. Doerr HW, Braun R, Munk K (1985) Human cytomegalo- virus infection : recent developments in diagnosis and epide- miology. Klin Wochenschr 63:241-251

507

10. Doerr HW, Rentschler M, Schleifler G (1987) Serologic de- tection of active infections with human herpes viruses (CMV, EBV, HSV, VZV) : diagnostic potential of IgA class and IgG subclass-specific antibodies. Infection 15:93-98

11. Engelhard D, Weinberg M, Or R, Shaked O, Naparstek E, Haikin H, Slavin S, Sarov I (1991) Immunoglobulins A, G, and M to cytomegalovirus during recurrent infection in recipients of allogenic bone marrow transplantation. J Infect Dis 163 : 628-630

12. Galea G (1990) The immune response to cytomegalovirus infection and serological methods used in its diagnosis. Med Lab Sci 47 : 297-303

13. Griffiths PD, Stagno S, Reynolds DW, et al. (1978) A longi- tudinal study of the serological and virological status of 18 women infected with CMV. Arch Virol 58 : 111-117

14. Henderly DE, Jampol LM (1991) Diagnosis and treatment of cytomegalovirus retinitis. J Acquir Immune Defic Syndr 4:$6-S10

15. Henderly DE, Freeman WR, Smith RE, Causey D, Rao NA (1987) Cytomegalovirus infection as the initial manifes- tation of the acquired immune deficiency syndrome. Am J Ophthalmol 103 : 316-320

16. Jabs DA, Enger C, Bartlett JG (1989) Cytomegalovirus re- tinitis and acquired immune deficiency syndrome. Arch Ophthalmol 107:75-80

17. Jacobson MA, Drew WL, Feinberg J, O'Donnell JJ, Whit- more PV, Miner RD, Parenti D (1991) Foscarnet therapy for gancyclovir-resistant cytomegalovirus retinitis in pa- tients with AIDS. J Clin Mierobiol 163:1348-1351

18. Marsano L, Perrillo RP, Wayne Flye M, Hanto DW, Spitzer ED (1990) Comparison of culture and serology for the diag- nosis of cytomegalovirus infection in kidney and liver trans- plant recipients. J Infect Dis 161:454-461

19. Miller H, McCulloch B, Landini MP, Rossier E (1989) Comparison of immunoblotting with other serological methods and virus isolation for the early detection of prirna- ry cytomegalovirus infection in allograft recipients. J Clin Microbiol 27: 267~2677

20. M/iller J, Weber B, Rabenau H (1991) Zytomegalievirus- diagnostik bei immunsupprimierten Patienten: Korrelation zwischen virologischen und klinischen Befunden. Innere Medizin 1 : 3-6

21. Ng VL, Gartner I, Weymouth LA, Goodman CD, Hopewell PC, Hadley WK (1989) The use of mucolysed induced spu- tum for the identification of pulmonary pathogens associat- ed with human immunodeficiency virus infection. Arch Pa- thol Lab Med 113:488~493

22. Pitchenik AE, Ganjei P, Torres A, Evans DA, Rubin E, Baier H (1986) Sputum examination for the diagnosis of Pneumocystis carinii pneumonia in the acquired immunode- ficiency syndrome. Am Rev Respir Dis 133:226-229

23. Rush JD, Ng VL, Hopewell PC, Hadley WK, Mills J (1989) Comparative recovery of eytomegalovirus from saliva, mu- colysed induced sputum, and bronchoalveolar lavage fluid from patients at risk for or with acquired immunodeficiency syndrome. J Clin Microbiol 27:2864-2865

24. Sarov I, Siqueira-Linhares M, Chardonnet Y, Levy E, Ay- mard M, Bosshard S, Nord E, Revillard JP (1981) Detection of specific IgA antibodies in serum of kidney transplant patients with reccurrent cytomegalovirus infection. Intervir- ology 15 : 228-234

25. Schacherer C, Braun W, Doerr HW (1988) Detection of cytomegalovirus in bronchial lavage and urine using a monoclonal antibody to an HCMC early nuclear protein. Infection 16: 288-292

26. Weber B, Enzensberger R, Doerr HW (1991) Zytomegalie und Schwangerschaft. TW Gyn/ikologie 1 : 5%63

27. Wunderli W, Auracher JD, Zbinden R (1991) Cytomegalo- virus detection in transplant patients: comparison of differ- ent methods in a prospective survey. J Clin Microbiol 29:2648-2650

Received : January 20, 1992 Accepted: March/2, 1992

Dr. B. Weber Abteilung ffir Medizinische Virologic Zentrum der Hygiene Universitfits-Kliniken Frankfurt/Main Paul Ehrlich-Strasse 40 W-6000 Frankfurt/Main 70, FRG