High prevalence of suspected HIV-associated dementia in adult Malawian HIV patients

INTRODUCTION

HIV-associated dementia (HAD) has been recognized as a common complication of advanced HIV infection in affluent settings. The clinical picture consists of cognitive, behavioural and motor dysfunctions. The prevalence was 15–20% in the pre-antiretroviral therapy (ART) era, but declined to around 5–10% after the introduction of highly active ART. ¹ In sub-Saharan Africa, where the main global burden of HIV disease exists, HAD has received relatively little attention. This may be due to the fact that HAD is less prevalent in Africa, but surveys from Uganda, Cameroon and Ethiopia dispute this. ² It is more likely that the difficulties in reaching the diagnosis of HAD in resource-limited settings explain the scarcity of data. HAD has no specific diagnostic test and the diagnosis is based on clinical criteria supported by neuropsychological testing and exclusion of other neurological conditions with neuroimaging and examination of the cerebrospinal fluid. ¹ Neuropsychological testing and neuroimaging are often not feasible in HIV clinics in sub-Saharan Africa, due to lack of time, expertise and equipment.

The International HIV Dementia Scale (IHDS) has been introduced in Africa as a cross-cultural, simple and appropriate screening tool for HAD in resource-limited settings that is suitable for illiterate persons. It has reasonable sensitivity but low specificity, ³ and has been used successfully in Ethiopia and Cameroon. ^4,5

HAD data are not collected routinely in the Malawi national ART scale-up programme and we are not aware of any published data from Malawi. Of the 12,000 patients initiated on treatment in our ART clinic since 2004, HAD was registered as the indication to start ART in only 16 cases. Coordinators of other large ART clinics in Malawi indicated that HAD was diagnosed with similar infrequency in Thyolo, Zomba, Lilongwe and Chiradzulu (E Augusto, A Chan, R Weigel, M Huckabee, personal communications). We performed a survey to acquire initial information about the prevalence of HAD and associated factors in Malawi and to introduce the IHDS in our practice.

METHODS

Malawi is a small country in sub-Saharan Africa with an HIV prevalence of 11.9% in adults aged 15–49 years. ⁶ A national ART scale-up programme began in 2004 and by the end of 2008 around 200,000 patients had been newly registered for ART. ⁷ This descriptive cross-sectional study took place at the large ART clinic of Queen Elizabeth Central Hospital in Blantyre, Malawi, which serves an urban and semi-urban population. At the time of the study, the clinic provided free ART to around 3500 persons.

The study was carried out on two study populations. The first consisted of patients who had come to determine eligibility for ART. The second included participants who had been taking ART for six months or longer. Other inclusion criteria were a positive HIV test and age of 18 years and above. The exclusion criteria were not being able to speak or hear, physical disabilities (i.e. upper limb amputation or paralysis) and severe medical or psychiatric illness interfering with study evaluation, active or known past central nervous system (CNS) infection and alcohol abuse (defined as >21 and 14 units/week for men and women, respectively). Every third patient seen by clinicians during the study period was approached for enrolment into the study.

A standard questionnaire was used to obtain demographic information. We derived a socioeconomic score from four yes/no questions used in the National Demographic Health Survey (tap water, electricity, iron roof, three meals per day) that generated a score from 0 to 4, which was categorized into two levels. We extracted the World Health Organization (WHO) clinical stage of HIV infection before the start of ART from the ART file or other available formal health record, recorded co-morbid medical and psychiatric conditions, including alcohol consumption and risk factors for vascular encephalopathy such as hypertension, diabetes and smoking, and measured weight and height.

The IHDS was used as described in detail elsewhere. ³ In brief, it consists of three elements. The first is memory assessment: four words are mentioned, after which the patient is asked to repeat and memorize them and then the patient has to recall the words 2–5 minutes later. The second is testing of motor speed by finger tapping: the patient is asked to open and close the first two fingers of the non-dominant hand as fast as possible, with ≥15 taps/5 seconds rated as normal. Lastly, psychomotor speed is assessed by instructing the patient to make alternate movements with the non-dominant hand as quickly as possible: clenching the hand in a fist, followed by putting the palm flat on the surface, followed by putting the ulnar side of the hand perpendicular to the flat surface, four sequences in 10 seconds being a normal result. Each element has a score from 0 to 4, with a maximum total of 12. Given the low specificity of the IHDS for HAD, a patient with a positive result (a score of 10 or less) should be considered to have suspected HAD. ⁵ The administration of the IHDS took place in English or the local language according to patients' preference.

We used Pearson's chi-square test to analyse categorical data. To determine factors associated with suspected HAD, we performed a binary logistic regression analysis. We included factors with a P value ≤0.2 in the univariate analysis as co-variates in the model. We calculated adjusted odds ratios and obtained P values based on the Wald test. SPSS software (version 12) was used for all analyses. The study was approved by the College of Medicine Research and Ethics Committee; written informed consent was obtained from all participants.

RESULTS

The study took place in July 2007. The first three authors were trained to use the IHDS by a neurologist within half an hour and were then able to explain and execute the procedure in around five minutes per patient. The only required equipment was a stopwatch. Of 220 adult patients who were approached for enrolment, 41 met exclusion criteria and 179 were enrolled. The majority of patients had clinically advanced HIV disease, 89.9% being in WHO stage 3 or 4. There was a female preponderance (65%) and the mean age was 36.7 years (range 18–65 years), both comparable to the ART population as a whole. Eight ethnic groups were represented: Lomwe (26%), Yao (23%), Ngoni (21%), Chewa (13%) and others (18%). Slightly more than 10% had a body mass index (BMI) less than 18.5 kg/m², indicating malnutrition. Forty-five patients had not yet started ART and 134 had been on ART for at least six months with a median duration of ART of 18 (range 6–79) months. The prevalence of suspected HAD, indicated by an IHDS test score of 10 or less, was 14.0% (95% confidence interval 8.9–19.1%). There was no significant difference between those not yet on ART and those on ART >6 months (15.6% versus 13.4%; P = 0.722). The median test score was 11.5 (range 7–12) and most points were lost in the word recall and psychomotor elements of the IHDS. Patients with suspected HAD were significantly more frequently men and had lower education in the univariate analysis (Table 1). Of the covariates included in the logistic regression model (gender, age category, education level category, socioeconomic category, BMI category), male gender and lower education level remained independently associated with suspected HAD (Table 1).

DISCUSSION

In our experience the IHDS was a feasible tool in the setting of a busy clinic that uses the public health approach to ART, requiring no specific expertise or expensive equipment, only very brief training and a fairly short time per test.

We found a considerable prevalence of suspected HAD; however, in the few other available studies from Africa even higher rates were recorded. For instance, in Uganda the HAD prevalence was 31% in 78 HIV-infected patients using more formal functional, neurological and neuropsychological evaluations and HAD was associated with low CD4 counts and advanced age. ⁸ In a study from Cameroon, the prevalence of suspected HAD determined with the IHDS in 204 patients not on ART was 21.1%, compared with 2.5% in non-HIV-infected controls. ⁵ The authors stress the usefulness of the IHDS as a screening tool and in a later report describe that advanced clinical stage and low CD4 count, as well as low haemoglobin were independent risk factors for suspected HAD. ⁹ A likely reason why HAD was less prevalent in our study is that most patients were on ART. Although we did not find a significant difference in the prevalence of suspected HAD among patients before ART versus those on ART for at least six months, our study had low statistical power to detect such a difference. Since ART can partially reverse and prevent HAD, ¹⁰ the prevalence of HAD in patients before the start of ART may be higher. Additionally, we may have taken a less selected sample than other studies, as we enrolled patients from a general ART clinic without specialist neurological facilities.

We have to consider some weaknesses of our study. We did not include a control group of non-HIV-infected persons, but used a cut-off level for the IHDS based on other African study results. The design of the study leaves open the possibility of bias, although the fact that age and sex were comparable to the ART clinic population as a whole suggests we took a representative sample. We were unable to determine the degree of immune suppression, since CD4 counts were not routinely collected in our clinic at the time of the study. We therefore cannot exclude the possibility that the association of male gender with suspected HAD is confounded by advanced immune suppression, since men generally present to ART clinics in Malawi at a later stage of disease than women. ¹¹ We did not establish the HIV clade in our patients. A recent study from Uganda has suggested that clade D virus may have a higher propensity to cause HAD than clade A; ¹² the paucity of data that is available from Malawi indicates that clade C is predominant in patients on ART, ¹³ yet its association with HAD is unknown.

In overcrowded ART clinics in sub-Saharan Africa consultation time is at a premium, thus the utility of the IHDS must be considered carefully. The IHDS can be used to select patients for more extensive neuropsychological tests and neuroimaging to confirm the diagnosis of HAD, but this will rarely be feasible in ART scale-up programmes using the public health approach, and no specific treatment with demonstrated efficacy is available for HAD, apart from ART. ¹⁴ First-line ART regimens in most ART programmes in sub-Saharan Africa are standardized, so it will not be possible to compose a regimen with the highest CNS penetration for patients with HAD, as is advised in affluent settings. ¹⁵ We therefore think that the main utility of the IHDS in sub-Saharan Africa will be in the identification of patients who would otherwise not be eligible for ART, reducing the chance that patients with HAD miss out on highly beneficial treatment. Should the IHDS become a routine tool to screen HIV-infected patients in sub-Saharan Africa, we need to learn more about the consequences of having a positive IHDS result; for instance, does this predict non-adherence and virological failure, or other unfavourable ART outcomes, such as loss to follow-up or death.

CONCLUSION

The IHDS is an appropriate tool to identify patients with suspected HAD in Malawi. In resource-limited settings, its main utility will be to identify HAD patients in need of ART. More study is required to determine the consequences of a positive IHDS result in relation to ART outcomes. The prevalence of suspected HAD in Malawian adults on ART appears to be considerable.