Post-traumatic stress disorder among persons with HIV who engage in heavy alcohol consumption in southwestern Uganda

Post-traumatic stress disorder (PTSD) is a significant worldwide health concern, and is highly prevalent in sub-Saharan Africa [1]. A recent meta-analysis showed that the pooled prevalence of probable current PTSD across 10 countries in sub-Saharan Africa was 22% [1]. These numbers are drastically higher than the estimated 4% prevalence of lifetime PTSD diagnosis reported in a 25-country study that included South Africa as the sole African country [2].

Sub-Saharan Africa is also disproportionately affected by the HIV/AIDS pandemic. Of the 38 million people estimated to be living with HIV/AIDS worldwide, over 70% live in sub-Saharan Africa [3]. While HIV has become a chronic illness, due to the wide spread roll out of antiretroviral therapy (ART), and mortality from AIDS greatly reduced [4], mental health issues remain common among persons living with HIV (PLWH) [5]. These issues include PTSD, depression, substance use disorders, eating disorders, and anxiety disorders, and they commonly co-occur [6, 7]. Probable PTSD occurs due to traumatic events experienced by the general population like serious accidents, natural disasters, political upheavals, and armed conflict, all which are common in sub-Saharan Africa [1]. PLWH in sub-Saharan Africa also may have increased risk of probable PTSD due to HIV-related traumatic events, such as intimate partner violence (IPV) and physical abuse [8].

Probable PTSD is associated with impaired quality of life, poor adherence with medical interventions, and high risk sexual behaviors [9, 10]. Among PLWH, probable PTSD may negatively affect HIV treatment outcomes [11, 12], adherence [13, 14], and retention in HIV care [15]. In addition, probable PTSD is further complicated by other common comorbid mental health disorders such as depression, as well as substance and alcohol use disorders [16]; among PLWH, probable PTSD commonly co-occurs with depression and alcohol use [17, 18]. Alcohol use is highly prevalent among PLWH [19] and may act as a coping mechanism for HIV-related stress [20, 21] and possibly probable PTSD.

While levels of probable PTSD in sub-Saharan Africa have been found to be high in several studies, there has been a high degree of heterogeneity in the results [1]. In addition, few studies have examined PTSD among PLWH in sub-Saharan Africa, and we are unaware of studies measuring probable PTSD among persons engaging in heavy alcohol use in this setting. Thus, we aimed to estimate the prevalence of probable PTSD in PLWH engaging in heavy alcohol use in Southwestern Uganda. Since untreated PTSD is associated with a host of negative consequences, and heavy drinking may be a response to PTSD which is not likely to be treated in low-income settings, we also sought to determine whether the level of alcohol consumption is associated with experiencing probable PTSD, and investigated other psycho-social factors as well. We conducted these investigations within a study of PLWH who engage in heavy alcohol consumption and who are co-infected with latent tuberculosis (TB) in Southwestern Uganda.

We screened 3293 persons with HIV from May 2018 through March 2020 (Fig. 1). Of those screened, 2611 people were eligible after the initial screening step; the main reasons for exclusion at this step were having a history of taking TB medications, a history of active TB, or taking nevirapine. At the next screening step which included urine EtG screening for recent alcohol use, and testing to exclude elevated liver enzymes, active TB, and pregnancy, the most common reason for exclusion was a negative EtG test (n = 1251); 1114 people remained eligible for PPD screening. Finally, 435 people (39%) who tested PPD-positive, were enrolled, and participated in the baseline survey. Four hundred twenty-one (421) participants were included in this analysis, as 14 had missing answers to the PTSD questions. Of the included participants, more than two-thirds (68.2%) were male and the median age was 40 (interquartile range [IQR] 32–47). The majority of the participants reported primary school education or less (78.6%), and 59.4% of the participants were either married or cohabiting. Participants reported their religious affiliation as Catholic (50.4%), Protestant (43%) and other denominations (6.7%). One in five of the participants (18.3%) reported having depressive symptoms (CES-D ≥ 16). More than half of the participants reported their general health status was excellent or very good (52.0%). The median AUDIT-C score was 6 [IQR 4–8]. Nearly 40% of participants reported having ever smoked tobacco in their lifetime, and 12.8% of participants reported ever using illicit drugs (Table 1). Khat (a chewed stimulant) was the most commonly used illicit drug, reported by 10%, marijuana was reported by 3.8%, kuba (a Ugandan informal word for a form of smokeless tobacco) by 3.1%, petrol sniffing by 1%; while no participants reported having ever used other illicit drugs.

In this analysis among PLWH who engage in heavy alcohol consumption in southwestern Uganda, we found that half of the participants surveyed reported experiencing a traumatic event and 21% met criteria for probable PTSD. We are aware of a handful of other studies that assessed probable PTSD among PLWH in low resource settings. The probable PTSD prevalence observed in this analysis was similar to the 19.6% prevalence of probable PTSD found in a study of PLWH in care in urban Uganda [34], lower than that observed in PLWH in a study conducted in Zimbabwe (55.3%) [35], and higher than that reported by a study among PLWH in post-conflict northern Uganda (8.3%) [9]. The prevalence we observed is quite similar to the 22% prevalence reported from a meta-analysis of data from 10 countries in sub-Saharan Africa that included some studies among PLWH, but did not calculate probable PTSD prevalence by HIV status [1]. The wide range of proportions of probable PTSD in these studies of PLWH and in the studies included in the above meta-analysis may be due to differences in scales used and differing populations. This study adds to the literature suggesting much higher rates of probable PTSD in low-resource settings in Sub-Saharan Africa compared to elsewhere [1, 36].

The level of alcohol use was not associated with probable PTSD in our study; this was consistent with the findings of a study among PLWH in a post-conflict area in northern Uganda [9]. However, we were only including people if they used alcohol at unhealthy levels, thus within this restricted sample, alcohol use severity and PTSD symptoms were not associated. But within the general population, alcohol use has been associated with probable PTSD [18, 37], and several theories have been put forward to explain this relationship. Alcohol use may lead to traumatic events that may cause probable PTSD, or may serve as self-medication to deal with symptoms of PTSD and therefore be a coping mechanism [20]. Our study population was comprised only of persons engaging in heavy drinking, so if alcohol is consumed to cope with probable PTSD up to only a certain level, e.g., a ceiling effect, such an effect might explain our failure to detect an association between level of drinking and probable PTSD. Another possible explanation for lack of association between level of drinking and probable PTSD in our sample could be underreporting of PTSD symptoms due to stigma and other barriers to admitting a mental health problem [38]. Underestimation of prevalence of mental health disorders is also another possible explanation for lack of significance between alcohol use and PTSD [39].

A notable finding was the association between depressive symptoms and probable PTSD. In our study, we found that having symptoms of depression was associated with probable PTSD; other studies of PLWH have reported similar associations [9, 40]. This association could be due to common overlapping symptoms between depression and probable PTSD [41]. Another possible explanation for this association could be the co-occurence of mental health disorders [35]. One study in South Africa that examined the prevalence and factors associated with probable PTSD found increased prevalence of depression among PLWH with probable PTSD compared to those without (29% versus 7%) [42]. Individuals with co-occurring probable PTSD and major depressive disorder tend to display more anxiety, depression and more PTSD symptoms and the occurrence is related to a more severe symptom presentation [43]. Co-existence of multiple mental disorders among PLWH is common and has been reported [35, 42]. Higher intrinsic religiosity was also associated with probable PTSD; other studies have reported religiosity as a coping mechanism for probable PTSD [44]. Those who had a history of lifetime smoking had higher odds of probable PTSD. Smoking has been found to be highly prevalent among patients with probable PTSD for its mood enhancement consequences and may serve as a coping mechanism for a cluster of negative affect, anxiety, and depression induced by probable PTSD [44‐47].

This study was embedded in a research study that had specific entry criteria, including requiring a purified protein derivative positive (PPD+) result and being on ART for at least 6 months, thus the generalizability may be limited. We did not measure the different specific life events that may have led to probable PTSD, so we have an incomplete picture of what may have precipitated the probable PTSD. The cross-sectional design limited the findings to associations only, and we are not able to establish cause and effect between probable PTSD and factors that were associated with probable PTSD, such as depression, religiosity, and smoking. Although the populations consisted of persons with confirmed heavy alcohol use, the level of drinking was self-reported and this may be subject to social desirability. Social desirability in reporting level of alcohol use has been demonstrated in this population [48], however we attempted to correct for this by including a social desirability scale.

All participants who were included in this study were confirmed to be persons who engage in heavy drinking through ethyl glucuronide testing. Drinking status can be particularly unreliable by self-report due to social desirability bias, so having an objective biologic measure is a strength for measuring this. Having four different sites, two of which were in a rural setting, with extensive clinical data from a relatively large care- seeking sample, was a strength.

While we found that probable PTSD is common among PLWH who engage in heavy alcohol consumption, we did not find a relationship between higher level of alcohol consumed and probable PTSD. The high level of probable PTSD and its association with depression suggests that there is a need for routine mental health screening in this population and more studies to assess coping mechanisms and design interventions for probable PTSD in this population.