Share this article on:

Induction and maintenance therapy of cytomegalovirus central nervous system infection in HIV-infected patients

Anduze-Faris, Béatrice M.a; Fillet, Anne-Marieb; Gozlan, Joëlc; Lancar, Rémid; Boukli, Narjisc; Gasnault, Jacquese; Caumes, Erica; Livartowsky, Joëlf; Matheron, Sophieg; Leport, Catherineg; Salmon, Dominiqueh; Costagliola, Dominiqued; Katlama, Christinea


Objective: To evaluate the efficacy and safety of the foscarnet–ganciclovir combination in induction therapy (IT) and maintenance therapy (MT) for cytomegalovirus (CMV) central neurological disorders in HIV-infected patients.

Design: An open pilot non-comparative multicentre study.

Methods: Thirty-one patients with acute CMV encephalitis (CMVe) (n = 17) or CMV myelitis (CMVm) (n = 14) during the era before highly active antiretroviral therapy (HAART) received intravenous IT with foscarnet 90 mg/kg plus ganciclovir 5 mg/kg twice a day followed by MT. The primary endpoint was clinical efficacy, assessed at the end of the induction phase.

Results: The foscarnet–ganciclovir combination in IT resulted in a 74% (23 out of 31 patients) clinical improvement or stabilization. Eight patients did not respond clinically. Side-effects leading to drug discontinuation occurred in 10 patients during IT. Among the 23 patients who qualified for the maintenance phase, CMV disease progressed in 10, with a median time to the first relapse of 126 days (range 64–264 days). Overall, the median survival time was 3 months [95% confidence interval (CI), 2–4 months].

Conclusion: The combination of foscarnet and ganciclovir can safely be used for CMV central nervous system (CNS) infection, with an improvement or stabilization in 74% of patients. Life-long MT with this combination is recommended as long as the immune system is profoundly impaired.

From the Departments of aMaladies Infectieuses et Tropicales, bVirologie, Hôpital Pitié-Salpétrière, Paris, France; cVirologie, dINSERM SC4, Hôpital Saint-Antoine, Paris, France; eMédecine Interne, Hôpital Bicêtre, Kremlin-Bicêtre, France; fMédecine Interne, Hôpital Béclère, Clamart, France; gMaladies Infectieuses, Hôpital Bichat, Paris, France; and hMédecine Interne, Hôpital Cochin, Paris, France.

Sponsorship: This work was supported by the Agence Nationale de Recherche sur le SIDA (ANRS 037).

Correspondence and requests for reprints to: Professor C. Katlama, Maladies Infectieuses, Hôpital Pitié-Salpétrière, 45–83 boulevard de l'Hôpital, Paris, France F75013.

This study was presented in part at the 6th European Conference on Clinical Aspects and Treatment of HIV Infection. Hamburg, 11–15 October 1997 [Abstract 400/B6] and the 5th Conference on Retroviruses and Opportunistic Infections. Chicago, 1–5 February 1998 [Abstract 263].

Received: 15 December 1999; accepted: 22 December 1999.

Back to Top | Article Outline


Until recently, cytomegalovirus (CMV) disease was a frequent and devastating opportunistic infection in HIV-infected patients with profound immunosuppression. Although retinitis is the most common form, CMV neurological manifestations have been reported in 18–28% of cases in autopsy series [1–5]. The detection of CMV in the cerebrospinal fluid (CSF) by means of the polymerase chain reaction (PCR) has markedly improved the diagnosis of CMV central neurological disease [6,7], allowing its use to select patients for therapeutic decision-making and clinical trials.

CMV encephalitis (CMVe) and myelitis (CMVm) are associated with a high mortality rate despite specific therapy [8–10]. The few available therapeutic studies mainly tested monotherapies, were retrospective, and involved a limited number of patients. In relapsing CMV retinitis the combination of foscarnet and ganciclovir has been shown to be more effective than foscarnet or ganciclovir alone in delaying CMV disease [11].

Given the poor clinical results obtained to date, the high mortality rate possibly caused by the limited penetration of foscarnet or ganciclovir in the brain [12,13] and the antiviral synergy of these two drugs in vitro [14], an open prospective trial of the foscarnet–ganciclovir combination was conducted in patients with acute CMVe or CMVm.

Back to Top | Article Outline

Patients and methods

Study design

This open prospective pilot study, conducted in 10 clinical centres in Paris, was designed to evaluate the efficacy and tolerance of the foscarnet–ganciclovir combination in HIV-infected patients with acute CMV central nervous system (CNS) infection. Induction therapy (IT) consisted of a 3 week course of foscarnet and ganciclovir given intravenously. At the end of IT, patients with neurological improvement or stable disease started life-long maintenance therapy (MT) with foscarnet–ganciclovir at half the IT dose regimen. The main primary endpoint was the clinical efficacy of the IT. Because of the severity of acute CMV neurological disease and poor outcome, this study was designed as a single arm study, which aims to evaluate the potentially optimal treatment for all patients.

The protocol was approved by the Institutional Ethical Review Board of Hôpital Pitié-Salpétrière in Paris.

Back to Top | Article Outline

Eligibility criteria

HIV-infected patients were eligible for the study if they had: acute neurological manifestations of either CMVe, defined as impaired mental functions, or CMVm defined as sensory-motor defects in the lower limbs. These manifestations had to be associated with the presence of CMV in the CSF detected by either culture or PCR reactions on two regions of the CMV genome, and an absolute CD4 cell count below 100 cells/mm3.

Patients were not eligible if they had a concomitant acute neurological disease, a previous episode of severe intolerance to foscarnet or ganciclovir, or as women if they were pregnant or breast-feeding. All the patients gave their written informed consent.

Back to Top | Article Outline

Study medication

IT consisted of foscarnet 90 mg/kg twice a day along with concomitant normal saline hydration (750–1000 ml twice a day) plus ganciclovir 5 mg/kg twice a day, given intravenously. The duration of IT was 3 weeks, but this could be extended up to 6 weeks if a slow clinical improvement occurred. MT consisted of foscarnet and ganciclovir given intravenously at a dose of 90 mg/kg once a day and 5 mg/kg once daily, respectively, along with normal saline.

Drug doses were adapted to body weight and serum creatinine levels. All concomitant medications were allowed, except for drugs potentially effective on CMV.

Back to Top | Article Outline

Follow-up and definitions

Baseline was defined as the first day of foscarnet–ganciclovir combination therapy. A complete history of both previous CMV disease episodes and treatments, and HIV-related conditions and therapies, was obtained at enrollment. A complete physical examination, including a neurological assessment, was performed. Patients were specifically interviewed on the possibility of extraneurological CMV manifestations. Fundoscopy and chest radiography were routinely performed, and endoscopic procedures were performed according to symptoms only.

Neurological signs and symptoms were assessed at baseline and every week during IT, then every 2 weeks during the first month of MT and every 4 weeks thereafter. The mini-mental status test was used to evaluate cognitive disorders, standard haematological and biochemical blood tests were performed at the same visits. The fundus was examined and CMV viraemia and viruria were assessed at baseline, at the end of IT, and every 2 months during MT. Lumbar puncture and complete CSF examination, including viral culture and two CMV PCR procedures were performed at baseline and at the end of IT. Neuroradiological assessments (cranial/lumbar computed tomography scan or magnetic resonance imaging according to clinical features) were made at baseline and were recommended at the end of IT.

At the end of the IT, the clinical response to therapy was scored as follows: complete response if all neurological signs and symptoms present at baseline resolved; partial response if signs and symptoms improved or stabilized; and progression if neurological manifestations worsened despite therapy.

During MT, progression of CMV disease was defined as either the recurrence or aggravation of neurological signs/symptoms, together with the detection of CMV in the CSF, or the onset of any active CMV extra-neurological localization. Follow-up ended in January 1998.

Back to Top | Article Outline

Virological methods

CMV viremia was assessed by the rapid detection of CMV antigen after rapid culture on confluent human fibroblasts, as previously described [7]. CMV was detected in the CSF by standard cell culture and qualitative PCR. Four hundred microlitres of CSF was inoculated on human fibroblasts, which were then observed for one month for the onset of CMV cytopathic effects. For qualitative PCR, two different CMV DNA amplifications were performed as previously described [15,16], in two virology laboratories. Results were considered to be positive when both PCR procedures were positive, indeterminate when one PCR was negative and the other positive, and negative when both PCR were negative.

CMV DNA was quantified retrospectively in CSF when samples were available, by means of competitive PCR in the CMV major immediate early gene [17], using a commercial kit (RS ELOSA CMV, Lambdatech, Namur, Belgium) allowing specific recognition of an internal standard and viral genomic DNA. Semi-quantitative results determined were obtained. When quantitative PCR was performed both at baseline and at the end of IT, the virological response was defined as `decreased viral load', `stable viral load' or `increased viral load'.

Back to Top | Article Outline

Statistical methods

The primary endpoints were the percentage of patients with clinical improvement at the end of IT, and survival. Secondary endpoints were the percentage of patients with CMV disease progression during MT, and tolerance.

Data on all patients who received at least one dose of the study drugs were analysed. The analysis was mainly descriptive in nature. Continuous variables were compared using the Wilcoxon non-parametric test. Survival rates were assessed using Kaplan–Meier estimates.

Back to Top | Article Outline


Study population

Thirty-one patients (25 men and 6 women) were enrolled in the study between July 1994 and October 1996; 17 had CMVe (55%) and 14 had CMVm (45%). The mean age was 38 ± 8 (range 28–57 years). The median CD4 cell count was 7/mm3 (range 0–84/mm3). In three patients, CMV CNS involvement was the first AIDS-defining event. The remaining 28 patients had had AIDS for a mean of 381 ± 338 days (range 12–1208 days). Nineteen patients (60%) had a previous history of CMV disease, with CMV retinitis in 18 and CMV colitis in one. The overall mean duration of CMV disease before CNS involvement was 159 ± 114 days (range 17–462 days). Among these 19 patients, 12 had received foscarnet before the study, 16 had received intravenous ganciclovir and seven had received oral ganciclovir, for median durations of 46 ± 53, 44 ± 50 and 43 ± 54 days, respectively. CNS involvement was the first manifestation of CMV disease in the remaining 12 patients (40%). At baseline, 10 out of 31 patients had active retinitis, and one also had severe CMV colitis.

Thirty of the 31 patients had received antiretroviral therapies for a median of 946 ± 600 days (range 11–2694 days). None of the patients had received protease inhibitors before the study. Only three patients received a combination of antiretroviral drugs including a protease inhibitor (n = 2) or a non-nucleoside analogue (n = 1) during the study period.

Back to Top | Article Outline

Baseline presentation

Baseline CMV neurological manifestations in the 31 patients are shown in Table 1. In the 17 patients with CMVe, the most common signs were cognitive dysfunctions: mental status was altered in all 17 patients, and was severely impaired (score < 24/30) in 70% of these subjects. By contrast, only three of the 14 patients with CMVm had cognitive dysfunctions. Gait disorders, pyramidal syndrome, sphincter dysfunction and sensitivity disorders prevailed in this latter group. Only 14 patients (45%) had fever.

Table 1

Table 1

CMV blood culture was positive in 17 out of 29 patients (58%). CSF findings are shown in Table 2. CSF was normal in 11 patients (35%). Seven patients with CMVe had elevated cellularity, with a median of 15 cells/ml (range 5–22 cells/ml), all of which were lymphocytes, whereas eight patients with CMVm had elevated cellularity, with a median of 53 cells/ml (range 10–2800 cells/ml), most of which were polymorphonuclear leukocytes. Nineteen patients had elevated CSF protein levels, with a median of 1.0 g/l in both groups. CMV CSF culture was positive in 7/17 patients (41%) at baseline. CMV CSF viral load was measured in 24 patients (see Table 1).

Table 2

Table 2

Baseline radiological assessments were performed in 29/31 cases. Eleven showed signs of CMV involvement, such as periventricular enhancement in 6/16 CMVe patients, or enhancement of the terminal cone and roots in 5/13 CMVm patients; eight patients had normal computed tomography or magnetic resonance imaging findings; 10 patients had non-specific abnormalities.

Back to Top | Article Outline

Induction therapy and efficacy

The clinical results of IT are shown in Table 2. Overall, 23 out of 31 patients (74%) had a clinical improvement at the end of IT, including two patients with CMVm, who had a complete response. The remaining eight patients either died during IT (n = 4) or had neurological progression at the end of IT (n = 4). The median duration of IT was 41 days (range 5–64 days). Twenty-one patients (68%) received the foscarnet–ganciclovir throughout the IT, whereas 10 patients discontinued one of the two drugs because of side-effects.

For 21 patients, CMV blood culture results were available both at baseline and at the end of IT: 13/21 were positive at baseline, and all but one were negative at the end of IT. Of the 19 patients who had CSF CMV PCR at the end of IT, only two were negative by PCR. In the eight patients whose CSF could be assessed for quantitative PCR both at baseline and at the end of IT, results showed a CMV viral load decreased in three patients (including the two negative results), stable in one and increased in four. Of the six patients who had CSF CMV culture both at baseline and at the end of IT, three were positive at baseline and all were negative at the end of IT.

Back to Top | Article Outline

Maintenance therapy and disease progression in 23 patients

Twenty-three patients (12 CMVe and 11 CMVm) received MT, consisting of foscarnet–ganciclovir in 13 cases (56%) and mainly of a monotherapy in 10. Overall, the median duration of MT was 74 days (range 10–484 days). Ten patients (43%) had CMV disease progression, with a median time from baseline to relapse of 126 days (range 64–264 days). CMV neurological relapse was the most common event. No statistically significant difference was found in the relapse rate between patients who received the foscarnet–ganciclovir combination and those who received monotherapy, although there was a trend towards a higher and earlier relapse rate in patients with CMVe (Table 3).

Table 3

Table 3

Back to Top | Article Outline

Tolerance and treatment discontinuation

Tolerance during IT and MT was assessed on the basis of drug discontinuation for side-effects. During IT, 10/31 patients discontinued at least one drug. As regards the maintenance period, there were 17 discontinuations by 14 of the 23 patients. In both cases most of the patients were re-challenged after resolution or improvement of the side-effects.

Back to Top | Article Outline


The overall median survival was 94 days (range 5–953 days), with survival rates of 77% after 6 weeks (95% CI, 62–92), 52% at 3 months (95% CI, 34–69) and 19% at 6 months (95% CI, 05–33). The median survival time among the 23 patients who received MT was 119 days. As of January 1998, two patients were still alive despite the discontinuation of all anti-CMV therapy and were receiving highly active antiretroviral therapy (HAART).

Multivariate analysis could not be used to identify factors predictive of relapse or death, owing to the small number of patients.

Back to Top | Article Outline


In the context of the poor prognosis of CMV acute neurological disorders and the suggested better efficacy of the combined regimen, we designed this open multicentre study to evaluate the clinical efficacy of combining forscarnet and ganciclovir in acute CMV-related CNS disorders in acute therapy and to provide information on long-term MT.

Overall, IT was clinically effective in 74% of the 31 patients presenting with acute CMVe or CMVm. This is the largest prospective series published to date. Comparison with published data is difficult because most studies were retrospective [10], and involved small numbers of patients [18–25]. However, with the exception of a few patients promptly treated at the onset of symptoms [26], CMV-related CNS diseases usually respond poorly to anti-CMV monotherapy, and survival is short [27]. Treatment failure may be explained by irreversible neurological damage caused by CMV, and by poor drug diffusion in the CNS. Whether the drug sensitivity of the infecting CMV strain influences clinical outcome remains to be evaluated. When this study was designed, we tried to overcome these problems by using rapid PCR diagnosis and early initiation of foscarnet–ganciclovir combination therapy known to be synergistic in vitro [14]. This therapeutic approach has previously been used in resistant CMV retinitis [11] and in a few cases of CMV neurological involvement [19,20,21,23,28].

In this study, the clinical presentation of CMV acute CNS disorders was equally distributed between CMVe (17 cases) and CMVm (14 cases); no stastistically significant difference was found in the efficacy of the foscarnet–ganciclovir combination in CMVe compared with CMVm, even though the two patients with complete response were both in the CMVm group and the four patients who died during IT were all in the CMVe group, suggesting that CMVe may be more difficult to control.

Most patients had a previous history of CMV disease, but CMV CNS disease was the first AIDS-defining event in three patients and the first CMV disease in 12/31 cases. As previously reported [7], none of the clinical manifestations were either consistently present or specific for CMV CNS involvement; fever was only observed in 45% of the patients at baseline; in addition, CMV blood culture was negative in 12/29 patients at baseline. Standard CSF examination was strictly normal in one-third of the patients, suggesting that the diagnosis of CMV CNS infection should be suspected in all patients with CD4 cell counts below 100/mm3 who present with central neurological signs, regardless of fever and CMV blood culture or results of standard CSF testing. All such patients should undergo virological analysis of their CSF. As previously reported [27,29,30], CMV was detected by CSF culture more frequently in patients with CMVm (five out of eight) than in patients with CMVe (two out of nine), confirming that qualitative PCR must always be performed when CMV neurological involvement is suspected.

In keeping with recent reports [31,32], quantitative PCR showed a broad range of viral load in CSF, with a trend towards higher values in CMVm (median between 2500 and 12 500 copies/10 μl) than in CMVe (median between 500 and 2500 copies/10 μl). Although contradictory results have been reported [33], quantitation of viral DNA in CSF has also been suggested potentially to predict or monitor the response to therapy [26,32,33]. In this study, no correlation was found between the baseline viral load and clinical response, as median quantitative PCR values were similar in patients with clinical improvement/stabilization and those who progressed. Similarly, no clear correlation was found between survival and changes in CSF CMV viral load between baseline and the end of IT in the eight evaluated patients, although the three patients who had a decrease in CSF viral load survived for 169, 181 and 544 days, i.e. longer than the median survival of the 23 patients who started MT (119 days).

Despite MT, 10 of the 23 patients (43%) experienced a CMV disease relapse, usually involving the CNS. This is very similar to the relapse rate observed in patients with retinitis despite MT [34].

The median time to the first relapse tended to be longer in the CMVm group (225 days) than in the CMVe group (101 days). At the time of progression, half the relapsing patients were receiving the foscarnet–ganciclovir combination therapy. In addition, the time to relapse did not differ between patients who received monotherapy and those who received combination therapy. Whether maintenance would be as efficient with one drug as with two drugs in preventing CNS relapse remains to be assessed. However, this combination therapy might be of value in preventing retinitis progression, as suggested by Weinberg et al. [11]. Indeed, only one case of retinitis relapse was observed, after 146 days, of the 14 patients who had a previous history of retinitis, of whom nine effectively received foscarnet–ganciclovir MT.

Overall, the foscarnet–ganciclovir combination was reasonably well tolerated, and was maintained in 65% of the patients throughout IT. Indeed a combination of two intravenous drugs may affect quality of life, with a higher toxicity rate and an increased duration of infusion. However, it must be emphasized that this issue may be considered more important during MT than during acute therapy, when immediate functional prognosis is the major concern. The overall median survival time was 94 days in this pre-HAART era study, confirming the severity of CMV CNS disease despite specific combination therapy. However, in a review of 28 published cases of patients with CMVe, most of whom received anti-CMV monotherapy, Arribas et al. [35] reported a median survival of 42 days (range 12–163 days). Only three patients received HAART during follow-up: one patient with CMVe had a survival of 405 days, two patients with CMVm were still alive at the cut-off date for analysis, 953 and 446 days after the diagnosis and discontinued anti-CMV drugs after immune restoration. The widespread use of combined antiretroviral therapy has led to a dramatic decrease in AIDS-related mortality and morbidity [36], and a lower incidence of opportunistic infections, including CMV disease, as a result of immune restoration [37,38].

However, in a prospective cohort study (PREDIVIR) of patients with baseline CD4 cell count below 200/mm3 before the introduction of protease inhibitor therapy, the incidence of CMV disease was 3.5 episodes per 100 patient-years [39]. The risk of CMV disease thus persists, and might increase in patients failing to respond to HAART. It is crucial that physicians rapidly recognize acute CMV neurological disease in order to consider early specific therapy. Presenting these results we consider that combined foscarnet–ganciclovir represents, in acute therapy of CMV CNS disease, a therapeutic regimen that can at best preserve the clinical prognosis. However, as for other CMV localizations, the optimal MT would be HAART in an era of effective CD4 cell immune restoration. Efforts to decrease HIV replication and induce immune restoration are likely to be the best way to control the progression of CMV disease.

Back to Top | Article Outline


The authors are indebted to all participants and their families, and would like to thank F. Conan and F. Angleraud for technical assistance.

The NEUROCMV Study Group consisted of the following clinical investigators, clinical research technicians, virologists and institutional pharmacists: L. Blum, D. Bollens, M. Bonmarchand, F. Boué, F. Bricaire, F. Brun-Vézinet, A. Certain, S. Chaput, J.P. Clauvel, J.P. Coulaud, A. De la Blanchardière, J.F. Delfraissy, J. Dormont, E. Dussaix, M. El Amrani, P. Faure, M.-H. Fievet, F. Fleur, J. Frottier, L. Gérard, V. Hazebroucq, S. Herson, N. Houhou, J.C. Imbert, L. Keros, S. Lariven-Jegat, P. Lebon, J. Livartowsky, S. Matheron, M.C. Meyohas, E. Oksenhendler, C. Olivier, O. Picard, N. Préaux, J.L. Prugnaud, Y. Quertainmont, M. Saillour, C. Scieux, D. Sicard, A.M. Taburet, P. De Truchis, F. Vachon, L. Vergely, C. Leport, J.L. Vildé.

Back to Top | Article Outline


1. McCutchan JA. Cytomegalovirus infections of the nervous system in patients with AIDS. Clin Infect Dis 1995, 20: 747 –754.
2. Fuller GN. Cytomegalovirus and the peripheral nervous system in AIDS. J Acquired Immune Defic Syndr 1992, 5 (Suppl. 1) : S33 –S36.
3. Roullet E, Assuerus V, Gozlan J. et al. Cytomegalovirus multifocal neuropathy in AIDS: analysis of 15 consecutive cases. Neurology 1994, 44: 2174 –2182.
4. Morgello S, Cho ES, Nielsen S, Devinsky O, Petito CK. Cytomegalovirus encephalitis in patients with acquired immunodeficiency syndrome: an autopsy study of 30 cases and a review of the literature. Hum Pathol 1987, 18: 289 –297.
5. Wilkes MS, Fortin AH, Felix JC, Godwin TA, Thompson WG. Value of necropsy in acquired immunodeficiency syndrome. Lancet 1988, 8602: 85 –88.
6. Fillet AM, Katlama C, Visse B, Camilleri S, Rogeaux O, Huraux JM. Human CMV infection of the CNS: concordance between PCR detection in cerebrospinal fluid and pathological examination [Letter]. AIDS 1993, 7: 1017 –1018.
7. Gozlan J, El Amrani M, Baudrimont M. et al. A prospective evaluation of clinical criteria and polymerase chain reaction assay of cerebrospinal fluid for the diagnosis of cytomegalovirus-related neurological diseases during AIDS. AIDS 1995, 9: 253 –260.
8. Mastroianni CM, Ciardi M, Folgori F. et al. Cytomegalovirus encephalitis in two patients with AIDS receiving ganciclovir for cytomegalovirus retinitis. J Infect 1994, 29: 331 –337.
9. Mussini C, Mongiardo N, Manicardi G. et al. Relevance of clinical and laboratory findings in the diagnosis of cytomegalovirus encephalitis in patients with AIDS. Eur Clin Microbiol Infect Dis 1997, 16: 437 –444.
10. Holland NR, Power C, Mathews VP, Glass JD, Forman M, McArthur JC. Cytomegalovirus encephalitis in acquired immunodeficiency syndrome. Neurology 1994, 44: 507 –514.
11. Weinberg DV, Murphy R, Naughton K. Combined daily therapy with intravenous ganciclovir and foscarnet for patients with recurrent cytomegalovirus retinitis. Am J Ophthalmol 1994, 117: 776 –782.
12. Hengge UR, Brockmeyer NH, Malessa R, Ravens U, Goos M. Foscarnet penetrates the blood–brain barrier: rationale for therapy of cytomegalovirus encephalitis. Antimicrob Agents Chemother 1993, 37: 1010 –1014.
13. Fletcher C, Sawchuk R, Chinnock B, de Miranda P, Balfour HH. Human pharmacokinetics of the antiviral drug DHPG. Clin Pharmacol Ther 1986, 40: 281 –286.
14. Manischewitz JF, Quinnan GV, Lane HC, Wittek AE. Synergistic effect of ganciclovir and foscarnet on cytomegalovirus replication in vitro . Antimicrob Agents Chemother 1990, 34: 373 –375.
15. Gozlan J, Salord JM, Roullet E. et al. Rapid detection of cytomegalovirus DNA in cerebrospinal fluid of AIDS patients with neurologic disorders. J Infect Dis 1992, 166: 1416 –1421.
16. Wolf DG, Spector SA. Diagnosis of human cytomegalovirus central nervous system disease in AIDS patients by DNA amplification from cerebrospinal fluid. J Infect Dis 1992, 166: 1412 –1415.
17. Demmler G, Buffone GJ, Scimbor CM, May RA. Detection of cytomegalovirus in urine from newborns by using polymerase chain reaction DNA amplification. J Infect Dis 1988, 158: 1177 –1184.
18. Kim YS, Hollander H. Polyradiculopathy due to cytomegalovirus: report of two cases in which improvement occurred after prolonged therapy and review of the literature. Clin Infect Dis 1993, 17: 32 –37.
19. Cohen BA, McArthur JC, Grohman S, Patterson B, Glass J. Neurologic prognosis of cytomegalovirus polyradiculomyelopathy in AIDS. Neurology 1993, 43: 493 –499.
20. Karmochkine M, Molina JM, Scieux C. et al. Combined therapy with ganciclovir and foscarnet for cytomegalovirus polyradiculomyelitis in patients with AIDS. Am J Med 1994, 97: 196 –197.
21. Peters M, Timm U, Schürmann D, Pohle HD, Ruf B. Combined and alternating ganciclovir and foscarnet in acute and maintenance therapy of human immunodeficiency virus-related cytomegalovirus encephalitis refractory to ganciclovir alone: a case report and review of the literature. Clin Invest 1992, 70: 456 –458.
22. Price TA, Digioia RA, Simon GL. Ganciclovir treatment of cytomegalovirus ventriculitis in a patient infected with human immunodeficiency virus. Clin Infect Dis 1992, 15: 606 –608.
23. Enting R, De Gans J, Reiss P, Jansen C, Portegies P. Ganciclovir/foscarnet for cytomegalovirus meningoencephalitis in AIDS [Letter]. Lancet 1992, 340: 559 –560.
24. Peters M, Schürmann D, Pohle H, Ruf B, Timm H. Combined and alternating ganciclovir/foscarnet in HIV-related cytomegalovirus encephalitis [Letter]. Lancet 1992, 340: 8825. 8825.
25. Jacobson MA, Mills J, Rush J. et al. Failure of antiviral therapy for AIDS-related cytomegalovirus myelitis. Arch Neurol 1988, 45: 1090 –1092.
26. Cinque P, Baldanti F, Vago L. et al. Ganciclovir therapy for cytomegalovirus infection of the central nervous system in AIDS patients: monitoring by CMV DNA detection in cerebrospinal fluid. J Infect Dis 1995, 171: 1603 –1606.
27. Miller RG, Storey JR, Greco CM. Ganciclovir in treatment of progressive AIDS-related polyradiculopathy. Neurology 1990, 40: 569 –574.
28. Cohen BA. Prognostic and response to therapy of cytomegalovirus encephalitis and meningomyelitis in AIDS. Neurology 1996, 46: 444 –450.
29. De Gans J, Portegies P, Tiessens G, Troost D, Danner SA, Lange JMA. Therapy for cytomegalovirus polyradiculomyelitis in patients with AIDS: treatment with ganciclovir. AIDS 1990, 4: 421 –425.
30. So YT, Olney RK. Acute lumbosacral polyradiculopathy in acquired immunodeficiency syndrome: experience in 23 patients. Ann Neurol 1994, 35: 53 –58.
31. Arribas JR, Clifford DB, Fichtenbaum CJ, Commins DL, Powderly WG, Storch GA. Level of cytomegalovirus DNA in cerebrospinal fluid of subjects with AIDS and CMV infection of the central nervous system. J Infect Dis 1995, 172: 527 –531.
32. Shinkai M, Spectror SA. Quantitation of human cytomegalovirus DNA in cerebrospinal fluid by competitive PCR in AIDS patients with different HCMV central nervous system diseases. Scand J Infect Dis 1995, 27: 559 –561.
33. Flood J, Drew WL, Miner R. et al. Diagnosis of cytomegalovirus polyradiculopathy and documentation of in vivo anti-CMV activity in cerebrospinal fluid by using branched DNA signal amplification and antigen assays. J Infect Dis 1997, 176: 348 –352.
34. Masur H, Whitcup SM, Cartwritht C, Polis M, Nussenblatt R. Advances in the management of AIDS-related cytomegalovirus retinitis. Ann Intern Med 1996, 125: 126 –136.
35. Arribas JR, Storch GA, Clifford DB, Tselis AC. Cytomegalovirus encephalitis. Ann Intern Med 1996, 125: 577 –587.
36. Palella FJ Jr, Delaney KM, Moorman AC. et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. :HIV Outpatient Study Investigators. N Engl J Med 1998, 338: 853 –860.
37. Autran B, Carcelain G, Li TS. et al. Positive effects of combined antiretroviral therapy on CD4+ T cell homeostasis and function in advanced HIV disease. Science 1997, 277: 112 –116.
38. Li TS, Tubiana R, Katlama C, Calvez V, Ait Mohand H, Autran B. Long-lasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV-1 disease. Lancet 1998, 351: 1682 –1686.
39. Salmon-Ceron D, Mazeron MC, Chaput S, et al. Evolution of CMV blood markers in HIV-infected patients receiving protease inhibitor (ANRS PREDIVIR cohort). 5th Conference on Retroviruses and Opportunistic Infections. Chicago, February 1998 [Abstract 752].

central nervous system; cytomegalovirus; encephalitis; foscarnet; ganciclovir; HIV; myelitis

© 2000 Lippincott Williams & Wilkins, Inc.