Effects of Different Integrase Inhibitors on Body Weight in Patients with HIV/AIDS: A Network Meta-Analysis

Background: Global antiretroviral therapy has entered the era of integrase strand transfer inhibitor (INSTI). Because INSTIs have the advantages of high antiviral ecacy, rapid virus inhibition, and good tolerance, they have become the rst choice in international acquired immunodeciency syndrome (AIDS) treatment guidelines. However, they may also increase the risk of obesity. There are differences in the effects of different INSTIs on weight gain in Human immunodeciency virus (HIV) infection / AIDS patients, but there is no evidence-based medical evidence. This study aimed to assess the effect of different INSTIs on body weight in HIV/AIDS patients. Methods: PubMed, Embase, Cochrane Library, China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature Database (CBM), China Science and Technology Journal Database, and Wanfang databases were searched by computer to screen the relevant literature on INSTI treatment of HIV/AIDS patients, extract the data on weight changes in the literature, and perform network meta-analysis using Stata16.0 software. Results: Eight articles reported weight changes in HIV/AIDS patients, and weight gain was higher after treatment with dolutegravir (DTG) than with elvitegravir (EVG) in HIV/AIDS patients, and the difference was statistically signicant [MD = 1.13, (0.18, 2.07)]. The network meta-analysis's consistency test results showed no overall and local inconsistency, and there was no signicant difference in the results of the direct and indirect comparison (P > 0.05). The rank order of probability was DTG (79.2%) > Bictegravir (BIC) (77.9%) > Raltegravir (RAL) (33.2%) > EVG (9.7%), suggesting that DTG may be the INSTI drug that causes the most signicant weight gain in HIV/AIDS patients. Conclusion: According to the literature data analysis, among the existing INSTIs, DTG may be the drug that causes the highest weight gain in HIV/AIDS patients, followed by BIC. BIC: Bictegravir; RAL: Raltegravir; ART: PRISMA: Reporting Items for Systematic Reviews Meta-Analyses; MD: mean difference; 95%CI: Condence interval; PI: protease inhibitor; NNRTI: non-nucleoside reverse transcriptase inhibitor; EFV: Efavirenz; MC4R: melanocortin four receptors; NA: Not available; F: Female; M: Male; RCT: Randomized controlled trial; Coef: coecient; Std.Err:

In the rst two years of ART treatment, patients with HIV/AIDS will have signi cant weight gain, which has become a recognized problem [11][12] . In the early ART stage, weight gain is an important manifestation of body rehabilitation, indicating the recovery of immunity and improving the survival rate in patients with HIV/AIDS [13][14][15][16][17] .

Literature screening and data extraction
The retrieved literature was initially screened by two investigators independently according to the inclusion and exclusion criteria and then cross-checked. The controversial literature was evaluated by the third party and uni ed by discussion. Two investigators extracted the relevant information of the included literature, including rst author, publication year, publication country, sample size, age, gender, and weight change.

Literature quality evaluation
Since the included studies were cohort studies or randomized controlled studies, the literature's quality was evaluated using the Newcastle-Ottawa scale (NOS) scale or the Jadad scoring scale.

Statistical Methods
The data were analyzed using STATA version 16.0 software. Measurement data were expressed as weighted mean difference (MD), and interval estimation was performed using a 95% con dence interval (CI) as an indicator of effect size. When the data extracted from the literature are brought into the stata16.0 version of the software for calculation, the node-splitting model in the software is used to compare the results of direct comparison with those of indirect comparison, to observe whether the two results are consistent, and then to make clear the consistency test results. If there was no statistical difference (P > 0.05), the consistent model was used for network meta-analysis of various drugs; if there was the statistical difference (P < 0.05), the source of nonconsistency was analyzed in detail. After comparing multiple interventions, rank probability ranking plots were used to rank the interventions and assess the drug that caused the most signi cant weight gain.

process and results of literature retrieval.
Four hundred ninety-seven related original articles were found in this network Meta-analysis, including 493 in English and 4 in Chinese, involving four intervention measures: DTG, RAL, EVG, and BIC. By carefully reading the titles and abstracts, screening the literature according to the inclusion and exclusion criteria, 65 articles were obtained and then excluded again by reading the full text. Finally, this study included eight articles [10,24−30] ( Fig. 1).

Basic characteristics and quality evaluation of the included literature
Among the eight included literature, there were 11,339 patients. The basic characteristics and quality evaluation of the included studies are shown in Table 1. The quality evaluation of the included studies showed that the overall quality of the literature was high.

Ranking of Probability for Weight Gain for Each Drug
The rank order of probability was DTG (79.2%) > BIC (77.9%) > RAL (33.2%) > EVG (9.7%), suggesting that DTG may be the drug that causes the most signi cant weight gain in HIV/AIDS patients (Fig. 3).

Consistency Test
Page 7/12 The whole study's inconsistency test results showed no signi cant difference between direct comparison and indirect comparison (P > 0.05). Therefore, there was no inconsistency between direct comparison and indirect comparison. The results of node analysis showed that there was no signi cant difference in direct and indirect comparison between BIC vs DTG, BIC vs EVG, DTG vs EVG, DTG vs RAL, and DTG vs RAL (P > 0.05), indicating that there was no local inconsistency (Table 3). INSTI is a class of drugs with high antiviral activity. In some studies, it was found that on the 10th day after treatment, HIV RNA levels in RAL, EVG, DTG and BIC groups could be reduced by 2.16 log 10 copies/ml [31] , 1.99 log 10 copies/ml [32] , 2.46 log 10 copies/ml [33] and 2.43 log 10 copies/ml [34] , respectively. However, after protease inhibitor (PI) and non-nucleoside reverse transcriptase inhibitor (NNRTI) treatment, the maximum amplitude of HIV RNA decrease was below 2 (1.29 ~ 1.99) log 10 copies/ml from 7 to 14 days [35][36][37][38] .
In many clinical trials of RAL, EVG, and DTG compared with NNRTI and PI drugs, INSTI showed good safety and tolerability, signi cantly reduced central nervous system adverse reactions compared with Efavirenz (EFV), and also signi cantly superior to PI drugs in terms of blood lipids and gastrointestinal adverse reactions [39][40][41] . The most common adverse reactions of INSTIs are nausea, vomiting, diarrhea, headache, and fatigue, and the severity of adverse reactions is only mild to moderate. Since the rst INSTI was applied in 2007, its overall acquired drug resistance rate is still low. The drug resistance rates of RAL, EVG, and DTG are 3.9%, 1.2%, and 0.1%, respectively [42] . In comparison, about 10-17% of treatment-naive HIV-infected individuals in high-income countries are resistant to NNRTIs, PI resistance is relatively rare, INSTI resistance transmission is even rarer, and DTG and BIC have a relatively higher resistance barrier [43] .
Given the high e ciency, safety, tolerability, and low drug resistance of INSTIs, ART has entered the era of INSTIs.However, with the widespread use of integrase, some studies have found that the application of integrase is associated with signi cant weight gain in infected individuals and even overweight in infected individuals. From the results of the probability ranking of the interventions in this study, it can be found that DTG is the drug that causes the most signi cant weight gain of patients among all INSTIs.The mechanisms by which different INSTIs contribute to weight gain are ambiguous and may be associated with multiple factors. An in vitro study reported that DTG in INSTI could inhibit the melanocortin four receptors (MC4R) [44] .MC4R plays a role in human energy homeostasis and correlates with human body weight. When the investigators knocked out the mice's MC4R gene, the mice showed severe obesity [45] . Therefore, after MC4R inhibition by DTG, the patient's body weight increased relatively more easily.Some scholars have pointed out that the difference in weight gain may be related to the inconsistent effect of different INSTIs on adipocytes. A study [46] found that the drug concentrations of different antiretroviral drugs in adipocytes are different, with higher DTG and EVG. An in vitro study [47] has demonstrated that EVG impairs adipocytes' metabolism, but RAL does not appear to damage adipocytes. Another hypothesis that may contribute to weight gain is that INSTIs affect the gut microbiota in HIV/AIDS patients [48] .
Studies by ElKamariV et al. [49] found that fatty acid-binding protein, as a marker of intestinal integrity, its level can be used as an independent predictor of weight gain and visceral fat gain in HIV/AIDS patients. Thus, the mechanism by which INSTIs cause weight gain in HIV/AIDS patients is not fully explained by a factor or mechanism.
An important reason for weight gain has attracted much attention because weight gain will increase the risk of non-AIDS-related diseases such as cardiovascular and cerebrovascular diseases in infected individuals. However, the risk of metabolic or cardiovascular diseases in HIV/AIDS patients cannot be completely and accurately predicted by the degree of obesity alone, which is due to the area of human fat distribution. Visceral fat or vascular content is bound to increase the risk of visceral or vascular diseases. Since the original literature included in this study did not analyze the differences between fat in peripheral or central regions of the body in HIV/AIDS patients, we could not assess the increase of fat in different body regions in HIV/AIDS patients INSTI.
Besides, whether some important metabolic parameters are correspondingly altered has not been effectively con rmed. It is expected that there will be subsequent studies on fat gain and metabolic parameters in different regions of the body in HIV/AIDS patients, and such results may be better able to guide the long-term health of HIV/AIDS patients.
The conclusions drawn from this study have certain reference value. However, since this original literature's quality was not as good as that of the randomized controlled trial in the included literature, the study subjects in the literature may have differences in basal body mass index, CD4 + T cell count, and patients' dietary habits. The correlation between these factors and weight change could not be excluded. Therefore, the conclusions drawn from this literature's data were not as strong as those of the randomized controlled trial. Besides, this study's conclusions are based on network Meta, and the strength of evidence is also weaker than the results of a direct comparison. Therefore, the interpretation of this study's conclusions should be cautious, and subsequent randomized controlled trials with rigorous design and large sample size are still needed to con rm further.

Conclusion
Based on data from the available literature, it is con rmed that DTG increases body weight most signi cantly in HIV/AIDS patients, while BIC is second only to DTG. However, there are still unresolved issues. It is unclear whether INSTI-based regimens cause lipohypertrophy (particularly an increase in visceral fat) or whether they increase the risk of cardiometabolic complications. As we continue to explore the road to ART, a more detailed description of these relevant studies that do not solve the problem has behaved crucially.

Competing interests
All authors declare that this research was conducted in the absence of any commercial or nancial relationships that could be construed as potential con icts of interest. Author Contribution RB,SL conceived, designed, and performed the analysis. LD veri ed the analytical methods. RB,SL wrote the paper and revised the manuscript for important intellectual content. HW revised the manuscript for important intellectual content. All authors discussed the results and contributed to the nal manuscript.