Spectrum of disease in HIV-positive patients presenting to a tertiary care hospital: a retrospective, cross-sectional review in Kumasi, Ghana

Although new HIV infections have declined significantly in western and central Africa in the last decade, these regions still account for 22% of total global infections [1]. The prognosis for persons living with HIV has improved dramatically since the advent of highly active antiretroviral therapy (HAART) [2], but these improvements have lagged behind in resource-poor countries [3]. Of the staggering 6.1 million people living with HIV in West and Central Africa, only 35% were accessing antiretroviral therapy in 2016 [1]. As a result, HIV-related morbidity and mortality remain unjustly high in this region [3–7]. Patients continue to present to inpatient care with WHO Class III and IV illness and undiagnosed or untreated HIV infection as a result of insufficient case finding, poor linkage to care and treatment failure or default [3, 4, 8–12]. This undoubtedly contributed to the 310,000 AIDS-related deaths that were reported in 2016 in western and central Africa [1]. In the nation of Ghana in the geographic region of West Africa, 230,000 adults were living with HIV in 2015 [13], but in this population there are currently limited morbidity and mortality studies of HIV-positive adults presenting to hospital.

Several studies in other nations have attempted to evaluate causes of hospitalization in HIV-infected patients and their associated mortality in resource-poor countries in order to better characterize these disparities [7, 11, 14–17]. The results are not surprising: most patients present with late stage HIV (WHO Classes III and IV), many have associated tuberculosis morbidity and resultant in-hospital mortality is high [3, 7, 14–16]. Regional differences in HIV co-morbidities have been noted within the African continent. For example, in East Africa, Cryptococcus neoformans infection continues to be a major burden of disease in immunocompromised patients [10]; however, the prevalence of cryptococcal antigenemia may be quite low in West Africa [18]. Thus, these studies are needed globally to identify the common HIV-related comorbidities on a regional basis. Epidemiological knowledge about disease presentations can inform better allocation of resources towards the diagnosis and treatment of common clinical presentations.

There are still advances in care that must be pursued in light of global targets in the HIV pandemic. The 90–90-90 goals advocate for substantial increases in case detection, treatment and follow-up care amongst this patient population, with goals of reaching 90% case diagnosis, 90% of patients diagnosed receiving antiretroviral treatment, and 90% of those on treatment achieving undetectable viral loads by the year 2020 [19]. In addition, the INSIGHT [20] and TEMPRANO [21] trials have led to the WHO’s 2015 recommendations for initiation of ART in all patients regardless of CD4 count [22].

In light of these recommendations and targets, our aim was to investigate the rate of known versus new HIV diagnosis as well as morbidity and mortality outcomes for HIV-positive patients admitted to the internal medicine department at the Komfo Anokye Teaching Hospital (KATH) in order to contribute to the development of strategies to improve care of individuals with HIV infection in Ghana.

As expected, pulmonary tuberculosis predominated among diagnostic findings, with 40.4% of admitted patients diagnosed with the condition during the course of their stay. Based on neurological findings on exam, 54 (22.0%) of patients were suspected to have an intracranial space occupying lesion (ICSOL). However, only 19 (35.2%) of these patients received a head CT scan. Fourteen of the 19 patients (73.7%) able to undergo head CT scan revealed radiographic evidence of toxoplasmosis (multiple, ring-enhancing lesions). Causes of mortality were varied. Of the 87 patients who died while admitted, 54.0% had pulmonary tuberculosis, while 29.9% had a clinical or radiological diagnosis of ICSOL. Another 29.1% were diagnosed with bacterial pneumonia. Other diagnoses are listed in Table 2.

The data collected from this study can aid in the development of strategies to reduce the morbidity and mortality of individuals with HIV infection in Ghana. It has identified the need for additional case finding and screening for HIV diagnosis. This study revealed a large number of newly diagnosed cases of HIV, with 40% receiving a new diagnosis during their current admission to the hospital. While some of these patients may have received a previous diagnosis that they chose not to disclose to the physician, there are likely a high number of new cases among patients that had not been recently screened. Thus, there is a need for increased HIV screening at the local level to enroll patients in treatment in order to achieve 90–90-90 targets [19] in the upcoming years.

The results of this study suggest that tuberculosis represents a major comorbidity (40.4%) amongst HIV-positive patients in this region of Ghana. This high prevalence of tuberculosis is similar to another study examining causes of morbidity at other sites in West Africa [4]. Additionally, neurological complaints commonly occurred in this patient population, with 22.0% of presenting patients thought to have an intracranial space occupying lesion on exam. Among these patients, less than half received CT imaging to confirm diagnosis representing a significant diagnostic barrier. This could contribute to the statistically significant association of neurological symptoms with mortality in the region. The possible reasons for not receiving CT imaging are multifactorial and would include high cost to the patient, lack of health insurance coverage for CT imaging, patient refusal, unavailability of CT imaging machines at the facility and mortality prior to receiving imaging during the inpatient hospital stay. Regional differences in causative organisms for meningitis were also highlighted, with five patients presenting with CSF findings consistent with tuberculosis meningitis (data not shown), while no patients had a positive test result recorded on serum or CSF for Cryptococcal infection. This finding exhibited the geographic variability of cryptococcal disease, which is common in eastern Africa [10] but does not appear to be prevalent in Ghana [13].

The results of this study suggest that resource allocation for HIV treatment in this region may be most useful in primary detection of HIV in order to prevent late patient presentation in stage 3 or 4 disease. Tuberculosis diagnosis and treatment must continue to be emphasized, as nearly half of patients diagnosed with pulmonary tuberculosis (47.5%) did not survive until discharge, while over half (54.0%) of all inpatient deaths were related to a tuberculosis diagnosis. Lastly, more resources must be allocated to diagnostic imaging for neurological complaints in HIV patients, as CT scanning is an essential component of diagnosis to differentiate symptoms concerning for intracranial lesions.

There are several potential weaknesses of the study. First, as data collection involved paper chart data retrieval, there are potential missed cases, as charts may have been removed from the ward before data collection was complete. Secondly, misdiagnosis is a possibility, as not all patients were able to undergo necessary investigations to confirm diagnoses, instead undergoing presumptive treatment. This also argues for resource funding from HIV aid groups for necessary diagnostic testing to guide treatment. Patients tested for HIV were screened using antibody testing, meaning some positive patients tested may have been missed during the window period. Lastly, there may have been patients admitted who did not undergo HIV testing or did not disclose their HIV-positive status who were nevertheless HIV-positive.

An additional benefit of this study involves the development of a low-cost, easily manageable database of de-identified patient information that could be used as a model for data collection in other resource-limited hospitals. This model could be expanded to collect data regarding any number of patient populations, not only the HIV-positive population. This study can serve as a foundation for further research to investigate the causes of the observed clinical presentations in this cohort of hospitalized patients—such as the role of patients’ level of access to HIV diagnostic testing and HAART and the prevalence of treatment failure, default, and non-adherence. Ultimately, such research will improve the lives of individuals in Ghana and continue to curtail the HIV epidemic.

In this population of HIV-positive patients presenting for care at a tertiary care center in Ghana, mortality rate was 35.5%, while 40.6% of patients received a new diagnosis of HIV on admission. Pulmonary tuberculosis was a major comorbidity, while neurological symptomology was a predictor of mortality. These findings argue for greater primary case detection, as well as allocation of resources for imaging and treatment of neurological conditions in HIV-positive patients.