HIV related hypochlorhydria does not appear to respond to anti-retroviral therapy in Zambian adults: a case control study

Introduction Human Immunodeficiency Virus (HIV) infection is associated with hypochlorhydria but the mechanism is unknown. The objective of this study was to determine effects of anti-retroviral therapy (ART) on gastric physiology as measured by validated markers. Methods We studied HIV infected individuals who were either ART-naïve or on treatment with undetectable viral loads. We measured H.pylori IgG antibodies, pepsinogen (PG) 1 and 2 levels and fasting gastrin-17 using Biohit GastroPanel®. Gastric antral biopsies and juice were obtained for histology and pH respectively. Also included were historical data from HIV negative participants (n = 72) in a previous study, for reference. Results We enrolled 84 HIV positive individuals with a median age 42 years (IQR 37-40 years). 55(66%) were female, 32(38%) were ART naïve, and 52(62%) were on ART. Hypochlorhydria (pH>4) was present in 48(57%) of the HIV positive and 18(25%) of the HIV negative individuals (OR 4: 95% CI 1.9-8.5, P<0.001) with no significant effect of ART (OR 0.9: 95% CI 0.3-2.3, P = 0.82). Hypochlorhydria was not associated with the serological detection of corpus atrophy using low PG 1:2 ratio (OR 2.1: 95% CI 0.5-10.2, P = 0.37) or GastroPanel® algorithm, (OR 0.7: 95% CI 0.01-60.1, P = 1.0). ART reduced the frequency of low PG 1:2 ratio (P = 0.001), but not the histological detection in the antrum of atrophy or non-atrophic gastritis. Conclusion ART use is associated with reduced serological evidence of corpus atrophy but has no effect on fasting pH, supporting earlier data that suggest that the mechanism of HIV-associated hypochlorhydria is multifactorial.


Introduction
Sub-Saharan Africa bears over two-thirds of the global Human Immunodeficiency Virus (HIV) burden. The prevalence of HIV in Zambia is 12.9% with 63% of these individuals being on antiretroviral therapy (ART) [1]. The use of ART has resulted in HIV infected individuals living longer and healthier lives and the classical opportunistic infections becoming less common [2]. The profile of health conditions in general and in HIV infection is beginning to change with a notable shift towards non-communicable diseases [2]. The gut plays an important role in HIV infection and it has been established that gut-associated lymphoid tissue continues to harbour HIV even when there is complete serological viral load suppression [3,4]. Hypochlorhydria (gastric pH greater than 4) is twice as common in HIV infection probably due to the presence of the virus [5,6]. Hypochlorhydria is typically a consequence of the loss of gastric parietal cell mass resulting from atrophy, a known risk factor for gastric cancer. There is no evidence to date of parietal cell loss in HIV infection [5]. In addition, the long-term consequences of persistent hypochlorhydria in HIV infection, including gastric cancer are unknown [7]. It is however known that hypochlorhydria predisposes to intestinal infections [5], micronutrient deficiencies [8] and other health outcomes. Over 15 years ago ART was rolled out to the general public in sub-Saharan Africa. Many HIV infected persons have benefited from this therapy, whose influence on gastric physiology remains uninvestigated. For the evaluation of gastric histology, endoscopy with biopsies is the gold standard but it is invasive and costly [9]. The option of serological assessment, "serological biopsy2 is therefore attractive.
The GastroPanel® (Biohit Helsinki, Finland) is a non-invasive diagnostic test for gastric pathology including four-biomarkers, pepsinogen 1 (PG 1), pepsinogen 2 (PG 2), Helicobacter pylori IgG (H.pylori IgG) antibodies and Gastrin-17. The utility of the GastroPanel® may prove to be a useful tool in detection of atrophic gastritis and gastric cancer risk assessment [10,11]. However, some studies have questioned its usefulness [12,13]. The aim of this study was to compare these markers of gastric physiology in HIV infected ART-naïve individuals and those who have taken treatment long enough to completely suppress HIV in blood. In addition, we included data from HIV negative individuals as a point of reference. To our knowledge, this tool has not been applied to the problem of hypochlorhydria in HIV infected individuals.

Methods
Participant enrollment: We enrolled participants from the University Teaching Hospital (UTH) in Lusaka, Zambia between October 2014 and July 2015. This was a case control study which included asymptomatic HIV infected individuals who were either ART-naïve as cases or on ART with full viral suppression ("ARTtreated") as controls. Only individuals above the age of 18 years and who had given written consent were included in the study.
Excluded were those with evidence of ART failure, use of gastric acid suppressing medication, presence of active gastrointestinal disease or significant co-morbidities such as chronic liver or kidney disease. For comparison, all proven HIV negative community volunteers from a previous study were included [14].

Results
We enrolled 84 clinically asymptomatic HIV positive participants. 32(38%) were ART-naïve and 52(62%) had been on ART for 2-14 years (median 10 years). The median age was 43 years (IQR 37-40 years); 66% were female. All individuals in the ART-treated group had undetectable viral loads and significantly higher CD4 counts than the ART-naïve group (medians 522 cells/μL and 297 cells/μL respectively, P=0.001). Our comparison group included 72 clinically asymptomatic HIV negative participants from a previous study (Table 1).  (Table 1). However, 37% of ART-naïve group had a low PG 1:2 ratio while in both the ART-treated and HIV negative groups the proportion was around 8% (Table 1). Using an unsupervised stepwise logistic regression, we found that being HIV positive increased the odds of having a low PG 1:2 ratio. However, individuals who were female, on ART and less than 40 years had significantly lower odds of having a low PG 1:2 ratio (Table 2). although the difference was not statistically significant (Table 3).

Discussion
In this study, we have shown that HIV-associated hypochlorhydria [5,17] is similar between ART-naïve and ART-treated individuals but not any of the serological indicators of corpus atrophy [18]. In To increase sensitivity, we obtained four antral biopsies using ordinary white light endoscopy. We were able to establish that the occurrence of gastric atrophy was not affected by ART use. We then evaluated the influence of ART on PG 1, PG 2 and gastrin-17 compared to histological diagnoses. PG 1 is produced in the corpus and its levels are reduced in corpus atrophy. When the corpus is inflamed without glandular loss, levels of PG 1 in blood increase [20]. PG 2 is produced in the corpus, antrum and duodenum.
Similar to PG 1 its levels will increase in non-atrophic gastritis [20].
In this study, we report consistent trends showing a significant increase in both PG 1 and 2 with non-atrophic gastritis and a decrease of both markers in atrophy. Similarly, Kitamura et al concluded that serum pepsinogen analysis could be used to predict the presence of H.pylori gastritis [21].
However, on histology only antral atrophy was analysed. We were therefore not surprised that that the decrease of pepsinogens in antral atrophy was insignificant. The third serological marker evaluated in this study was fasting gastrin-17. The G cells of the antrum produce basal gastrin-17 [20]. It is expected that levels of gastrin-17 would decrease when the antral mucosa is atrophic.
Paradoxically our results did not show any association between antral atrophy and gastric-17. In addition, we did not find any effect of HIV disease progression on the levels of fasting gastrin-17 (which would be indicative of antral inflammation) as determined by some other investigators. The levels of serum pepsinogens were similar regardless of the presence of H.pylori on histology. Histology is specific for H.pylori yet its sensitivity is hampered by the patchy colonization of the stomach [22]. The median age of participants in the ART-naïve group was significantly lower than in the ART-treated or the HIV negative groups. As the incidence of gastric atrophy increases with advancing age, we would expect the occurrence of low PG 1:2 ratio to be less in the ART-naïve group. However, our results showed a significantly higher proportion of low PG 1:2 ratio in the younger ART-naïve group, an effect that seem to be reversed on ART. We therefore concluded that HIV infection could be directly or indirectly influencing the production of pepsinogens, through a mechanism similar to that causing hypochlorhydria. Several investigators have published reports supporting the use of "serological biopsy" to evaluate gastric physiology, but this has not yet been examined in HIV infection. It is clear that HIV is associated with physiological changes and therefore any such investigative strategies would warrant specific validation. We therefore explored the utility of the GastroPanel® kit for detection of gastric atrophy in HIV infection. To analyse our data, we used both the algorithm for

Conclusion
We have shown that ART does not reverse HIV associated hypochlorhydria. The mechanism of low acid output in HIV infection is yet to be established.

What is known about this topic
 HIV infection is associated with hypochlorhydria;  Anti-retroviral therapy can suppress the viral load to undetectable levels;  Use of the GastroPanel® has not been validated in HIV infection.

What this study adds
 Anti-retroviral therapy does not reverse HIV associated hypochlorhydria;  The occurrence of gastric atrophy is not affected by antiretroviral therapy;  Anti-retroviral therapy reduces the occurrence of low pepsinogen 1:2 ratio.