SEARCH

SEARCH BY CITATION

Keywords:

  • Cryptococcus;
  • HIV;
  • Africa;
  • epidemiology;
  • cryptococcal antigen
  • Cryptococcus;
  • VIH;
  • Afrique;
  • épidémiologie;
  • antigène cryptococcal
  • Criptococo;
  • VIH;
  • África;
  • epidemiología;
  • antígeno

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgement
  5. References

Objectives  To determine the prevalence of cryptococcal antigenaemia in a clinic population with advanced HIV infection, with a view to giving antifungal therapy to those testing positive.

Methods  Serum samples from adults with CD4 count <100 cells/mm3 presenting to a large HIV clinic in Kumasi, Ghana, were tested retrospectively for cryptococcal antigenaemia using a latex agglutination assay, and clinical and demographic data extracted from case notes.

Results  Of 92 samples tested, two were positive thus giving a prevalence of 2% (95% CI, 0–5.2%).

Conclusions  The prevalence of cryptococcal antigenaemia in patients with advanced HIV infection enrolling in an antiretroviral programme appears to be low in Kumasi, suggesting that the value of routine testing of outpatients diagnosed with advanced HIV infection may be limited in this population.

Objectifs:  Déterminer la séroprévalence de l’antigène cryptococcal dans une population clinique avec une infection VIH avancée, en vue d’administrer un traitement antifongique à ceux trouvés séropositifs.

Méthodes:  Des échantillons de sérum d’adultes avec des taux de CD4 inférieurs à 100 cellules/mm3 se présentant dans une grande clinique VIH à Kumasi, au Ghana, ont été analysés rétrospectivement pour l’antigène cryptococcal en utilisant un test d’agglutination au latex et les données cliniques et démographiques ont été extraites des notes des cas.

Résultats:  Sur 92 échantillons testés, 2 étaient positifs soit une prévalence de 2% (IC95%: 0 à 5,2).

Conclusions:  La séroprévalence de l’antigène cryptococcal chez les patients inscrits avec une infection VIH avancée dans un programme antirétroviral apparaît faible à Kumasi, ce qui suggère que la valeur d’un test de routine chez les patients ambulatoires avec un diagnostic de l’infection VIH avancée pourrait être limitée dans cette population.

Objetivos:  Determinar la prevalencia de la antigenemia para criptococo en una población clínica con una infección avanzada por VIH, con el fin de proveer terapia antifúngica a aquellos que dan positivo.

Métodos:  Muestras de suero de adultos que se presentan en una clínica para VIH en Kumasi, Ghana, y que tienen un recuento de CD4 de menos de 100 células/mm3. Se les realizó, de forma retrospectiva, una prueba de antigenemia para criptococo mediante aglutinación de partículas de látex, y se extrajeron datos clínicos y demográficos de las historias clínicas.

Resultados:  De las 92 muestras evaluadas, 2 eran positivas, dando una prevalencia del 2% (IC 95%, 0–5.2%).

Conclusiones:  La prevalencia de antigenemia para criptococo en pacientes con una infección avanzada por VIH que participan en un programa de antirretrovirales parece ser baja en Kumasi, lo cual sugiere que el valor de realizar la prueba a los pacientes atendidos a través de servicios externos, diagnosticados con una infección avanzada por VIH, es limitado en esta población.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgement
  5. References

Cryptococcal meningitis (CM), a fungal infection caused by Cryptoccoccus neoformans, has emerged as a leading cause of mortality in HIV-infected patients in the developing world (Okongo et al. 1998; Corbett et al. 2002; French et al. 2002; Castelnuovo et al. 2009). In sub-Saharan Africa, CM accounts for 20–50% of this early mortality, with an estimated half million of AIDS-related deaths per year (Park et al. 2009). Studies in Uganda and Tanzania have shown that 80–90% of patients with CM had CD4 counts of <100 cells/mm3 (French et al. 2002; Kisenge et al. 2007). Cryptococcal antigenaemia is detectable a median of 22 days before the onset of symptoms (French et al. 2002) and has been shown to be 100% sensitive for predicting the development of CM in the first year of antiretroviral therapy (ART) (Jarvis et al. 2009), as well as being associated with both CM and mortality in two Ugandan studies (Castelnuovo et al. 2009; Meya et al. 2010). Cryptococcal antigen (CRAG) tests on serum are highly sensitive and specific and have been validated for use in the HIV-infected population (Temstet et al. 1992; Tanner et al. 1994; Lara-Peredo et al. 2000). Screening for subclinical or asymptomatic infection by a serum CRAG assay in patients with advanced HIV infection, and giving antifungal therapy to those testing positive, may prevent the development of CM; the use of fluconazole in patients with asymptomatic cryptococcal antigenaemia was associated with an odds ratio of survival of 26.2 in one study (Meya et al. 2010).

The prevalence of cryptococcal antigenaemia in patients presenting with advanced HIV infection has been estimated to be between 7% and 13% in Uganda, South Africa and Thailand (Tassie et al. 2003; Jarvis et al. 2009; Meya et al. 2010; Pongsai et al. 2010), but is still unknown in many parts of Africa, including West Africa. The only population-based estimates of the burden of cryptococcal disease in this region are based on data extrapolated from provider-based studies (Park et al. 2009), and CM is infrequently diagnosed in patients admitted with advanced HIV infection in Kumasi. We conducted a retrospective study in a large government hospital in Ghana on stored serum samples to estimate the prevalence of cryptococcal antigenaemia in outpatients presenting with advanced HIV infection and to study the effect of cryptococcal antigenaemia in terms of short- to medium-term mortality, response to ART and clinical features of subsequent cryptococcal disease.

The approval of the Committee on Human Research and Ethics at KNUST was obtained for this study. Serum samples of adult HIV-positive patients, diagnosed from April 2008 to May 2009 and who attended the HIV clinic at Komfo Anokye Teaching Hospital (KATH), were collected and stored. Patients who had an initial baseline CD4 < 100 cells/mm3 and had a serum sample taken within 6 months of diagnosis (or within 3 months of starting ART) were eligible for inclusion in this study. Demographic data, ART (if applicable), body mass index (BMI), CD4 counts, HIV stage and documented clinical events or clinical signs suspicious of cryptococcal disease were recorded. Stored serum samples were tested using Latex-Cryptococcus Antigen Detection System (Immuno-Mycologics, USA), according to the manufacturer’s instructions, both in Kumasi and later in London (for confirmation), with CRAG titres estimated on positive samples.

Ninety-two patients with a serum sample available were eligible for the study, representing over 80% of all patients who presented with a CD4 < 100 cells/mm3 in this time period. The median time from a positive HIV test to obtaining the sample used for CRAG testing was 2 days (IQR 1–6), with 78% of samples obtained within 1 week of diagnosis of HIV infection. Only two serum samples were positive for cryptococcal antigenaemia, one positive up to a dilution of 1:16 and the other only positive on neat serum. The prevalence of antigenaemia in this population is therefore estimated to be 2% (95% CI, 0–5.2). The first patient presented to the clinic in 2008 with a baseline CD4 count of 31 cells/mm3. The patient was a 58-year-old man with a BMI of 16 and WHO Stage 1 disease at diagnosis. There were no clinical symptoms or features of cryptococcal infection at presentation or follow-up. He defaulted follow-up after his second clinic visit and did not start ART. The second patient with positive CRAG on neat serum only, so not a clear positive, was a 40-year-old woman who presented to the clinic in 2008 with a CD4 count of 4 cells/mm3 and pulmonary tuberculosis. She was started on ART 6 months after diagnosis and has not defaulted from follow-up. Because of the low prevalence rate of cryptococcal antigenaemia, no meaningful comparison between demographic and clinical features of the CRAG-positive and CRAG-negative patients was feasible.

Table 1 shows the baseline characteristics of the study population. Demographic and other features of this cohort were similar to the overall group of patients presenting with a CD4 count below 100 cells/mm3. The median CD4 count at diagnosis was 28 cells/mm3 (IQ range 8–54), whilst most patients were classified as WHO Stage 3. Comprehensive clinical data were available for 56 patients, of whom 28 presented with only one symptom at diagnosis and 28 with two or more when first seen in clinic. Symptoms recorded were almost entirely non-specific e.g. weight loss. Fifty-three patients defaulted follow-up in clinic, mostly within the first 6 months of follow-up, and although precise data on mortality were not available (many patients travelled long distances to attend clinic), it is likely many of these patients died shortly after diagnosis. In this group, only 18 were started on ART prior to defaulting, with a median delay of 36 days (IQ range 21–58) from diagnosis. There were 39 patients who continued to attend clinic, and who were all started on ART with a median delay of 33 days from diagnosis (IQ range 28–50). Interestingly, a high proportion presented with a WHO stage of 1/2 despite having a CD4 < 100 cells/mm3.

Table 1.   Baseline characteristics of study population
 CharacteristicInterquartile range
  1. Median values shown unless stated otherwise.

Gender
 Male (%)38 (41%) 
 Female (%)54 (59%) 
Age4033–46
Body Mass Index (kg/m2)20.219.7–21.2
CD4 count (cells/mm3)288–54
TB diagnosed at presentation13 (14%) 
WHO stage (N)58 
 I12 (21%) 
 II11 (19%) 
 III26 (45%) 
 IV9 (15%) 
Non-defaulters
 N39 (42%) 
 Time (days) to starting antiretroviral therapy (ART) from diagnosis3328–50
Defaulters
 N53 (58%) 
 Median number of days of follow-up1678–374
 Number started on ART prior to defaulting18 (34%) 
 Time (days) to starting ART from diagnosis3621–58

This study has demonstrated a low prevalence of cryptococcal antigenaemia in outpatients presenting with advanced HIV infection in Ghana. If the true prevalence in this population is significantly lower than that observed in other populations in sub-Saharan Africa (given the upper 95% confidence limit of 5.2% may overlap with the lower limit from other studies), it remains unclear why the prevalence is lower. Although the sample size tested in this cohort was smaller than the three other populations studied in Uganda and South Africa (Tassie et al. 2003; Jarvis et al. 2009; Meya et al. 2010), the mean CD4 count in our cohort was substantially lower, which makes the apparent lower prevalence even more surprising. One possibility to explain this difference is selection bias, such that more patients in Kumasi with cryptococcal infection died before either being tested for HIV or attending the clinic. It remains unclear how significant this effect may have been given that referral processes and HIV testing rates appear to be similar in Ghana to Uganda and South Africa. However, it is possible that more patients with advanced HIV infection who were referred for a HIV test or admitted to hospital died of cryptococcal disease before having a HIV test and attending clinic. Unfortunately, as the CRAG test was not routinely available for patients admitted to the hospital, it was impossible to obtain data on rates of cryptococcal antigenaemia in patients admitted with advanced HIV infection or AIDS; clearly it would be useful to obtain these data in a future study.

The cost of preventing one death through CRAG screening (and fluconazole treatment) has been estimated to be $266 in a Ugandan cohort with a prevalence of asymptomatic antigenaemia of 13.5% (Meya et al. 2010). The cost effectiveness of such a screening strategy significantly diminishes once the prevalence of ‘asymptomatic’ antigenaemia falls below 5%, with an estimated cost of preventing one death being over $500 (Meya et al. 2010). Extrapolating the same cost-benefit analysis from Uganda to Ghana, and assuming costs of CRAG assays and fluconazole therapy are similar, the cost of preventing one death in Ghana, with a prevalence of antigenaemia between 0% and 5%, would range from $325 to over $3000. Hence even only selecting those with advanced HIV infection, if these costs are comparable, screening may not be economically viable. On the other hand, if costs of CRAG tests are lower than the $16/test assumed by Meya et al. (2010) or alternative strategies for reducing costs such as testing pooled samples are adopted, or the costs of antifungal drugs fall substantially, it may prove cost effective to screen patients with advanced HIV infection in Ghana.

This study also showed that most patients had to wait more than a month before starting ART, although in a proportion, the delay was due to their starting anti-tuberculous therapy. The high default rate (and presumed mortality) in this specific population suggests a need to review the HIV care programme in Ghana, to enable patients presenting with low CD4 counts to start ART sooner.

Acknowledgement

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgement
  5. References

Financial support for this study was provided by the South Tees Hospitals NHS Foundation Trust Academic Foundation Year Programme.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgement
  5. References
  • Castelnuovo B, Mannabe YC, Kirraga A, Kamya M, Easterbrook P & Kambugu A (2009) Cause-specific mortality and the contribution of immune reconstitution inflammatory syndrome in the first 3 years after antiretroviral therapy initiation in an urban African cohort. Clinical Infectious Diseases 49, 965972.
  • Corbett EL, Churchyard GJ, Charalambos S et al. (2002) Morbidity and mortality in South African gold miners: impact of untreated disease due to human immunodeficiency virus. Clinical Infectious Diseases 34, 12511258.
  • French N, Gray K, Watera C et al. (2002) Cryptococcal infection in a cohort of HIV-1 infected Ugandan adults. AIDS 16, 10311038.
  • Jarvis JN, Lawn SD, Vogt M, Bangani N, Wood R & Harrison TS (2009) Screening for Cryptococcal Antigenaemia in patients accessing an antiretroviral treatment program in South Africa. Clinical Infectious Diseases 48, 856862.
  • Kisenge PR, Hawkins AT, Maro VP et al. (2007) Low CD4 count plus coma predicts cryptococcal meningitis in Tanzania. BMC Infectious Diseases 7, 39.
  • Lara-Peredo O, Cuevas LE, French N, Bailey JW & Smith DH (2000) Cryptococcal infection in an HIV positive Ugandan population. Journal of Infection 41, 195.
  • Meya DB, Manabe YC, Castelnuovo B et al. (2010) Cost-effectiveness of serum cryptococcal antigen screening to prevent deaths among HIV-infected persons with a CD4+ cell count < or = 100 cells/microL who start HIV therapy in resource-limited settings. Clinical Infectious Diseases 51, 448455.
  • Okongo M, Morgan D, Mayanja B, Ross A & Whitworth J (1998) Causes of death in a rural, population-based human immunodeficiency virus type 1 (HIV-1) natural history cohort in Uganda. International Journal of Epidemiology 27, 698702.
  • Park BJ, Wannemuehler KA, Marston BJ et al. (2009) Estimation of the current global burden of cryptococcal meningitis among persons living with HIV/AIDS. AIDS 23, 525530.
  • Pongsai P, Atamasirikul K & Sungkanuparph S (2010) The role of serum cryptococcal antigen screening for the early diagnosis of cryptococcosis in HIV-infected patients with different ranges of CD4 cell counts. Journal of Infection 60, 474477.
  • Tanner DC, Weinstein MP, Fredorciw B, Joho KL, Thorpe JJ & Reller LB (1994) Comparion of commercial kits for detection of cryptococcal antigen. Journal of Clinical Microbiology 32, 16801684.
  • Tassie JM, Pepper L, Fogg C et al. (2003) Systematic screening of cryptococcal antigenemia in HIV-positive adults in Uganda. Journal of Acquired Immune Deficiency Syndromes 33, 411412.
  • Temstet A, Roux P, Poirot JL, Ronin O & Dromer F (1992) Evaluation of a monoclonal antibody-based latex aggluntination test for the diagnosis of cryptococcosis: comparison with two tests using polyclonal antibodies. Journal of Clinical Microbiology 30, 25442550.