Identifying risk factors for late HIV diagnosis and survival analysis of people living with HIV/AIDS in Iran (1987–2016)

Background Late-diagnosis of HIV is a major challenge for the control and prevention of AIDS in the world. The present study aimed to specify factors associated with the late diagnosis of HIV in Iran from 1987 to 2016. Methods In this retrospective cohort study, data for 4402 diagnosed HIV/AIDS patients were extracted from 158 behavioral disease counseling centers of 31 Iranian provinces. We defined late diagnosis as having a CD4 count less than 350 within 3 months after diagnosis. Multiple logistic regression analysis was used to determine the factors influencing late diagnosis. Moreover, we used multivariate Cox regression to assess the association of these factors with the patients’ survival. Results In this study, the prevalence of late diagnosis among the patients was 58.2%. People aged 50 years and over (adjusted OR = 3.55), transmission through blood transfusion (adjusted OR = 2.89), co-infection with tuberculosis (adjusted OR = 2.06), and male gender (adjusted OR = 1.38) were the strongest predictors for late diagnosis of HIV. On the other hand, baseline CD4 (adjusted HR = 2.21), people aged 50 and over (adjusted HR = 1.81), male gender (adjusted HR = 1.76), being a widow (adjusted HR = 1.68), people with unknown transmission way (adjusted HR = 18.24), people who inject drugs (adjusted HR = 1.87), diagnosis at previous years (adjusted HR = 2.45) and co-infection with tuberculosis (adjusted OR = 1.77) significantly associated with the survival of patients. Conclusion The prevalence of late diagnosis is high among Iranian HIV/AIDS. The risk factors of late diagnoses include being males and aged 50 years and over, transmission through blood transfusion, and co-infection with tuberculosis. Therefore, implementation of screening programs for early diagnosis of HIV these high risk groups is recommended to Iranian health providers and policymakers.


Introduction
Acquired immunodeficiency syndrome (AIDS) is a potentially life-threatening condition caused by the human immunodeficiency virus (HIV). Globally, about 40 million people are living with this virus, and every year 1.7 million new cases are added to existing cases. Moreover, yearly 700,000 people die from AIDS in the world [1].
Recommended treatment for AIDS is called antiretroviral therapy (ART), which involves taking a combination of HIV medicines [2]. ART aims to reach viral load suppression of HIV in the body, that is, ART reduces the viral load of HIV in the blood to a very low level, and promotes immune system performance, and prevents diseases [3]. Furthermore, ART reduces the probability of HIV transmission to other people when viral load is very high. Despite ART does not cure AIDS, it assists the patients to have a longer and healthier life [4]. Pre-requirement for successful treatment of AIDS is early diagnosis of HIV in the suspected people. Early HIV infection is defined as being diagnosed 6 months after initiation of HIV infection [5]. Despite the importance of early diagnosis of HIV in patients' treatment, the studies show that late diagnosis has been remained a major problem for the control and prevention of AIDS in the world [6]. Late diagnosis of HIV adversely affects individuals and society by reducing the effectiveness of antiretroviral therapy (ART) and subsequently increasing the risk of morbidity and mortality in the patients. On the other hand, this late diagnosis increases the probability of transmission and spread of the disease in societies, and as a result, control of the disease becomes harder [7,8].
Studies identified several factors affecting the late diagnosis of HIV. They demonstrated that male gender and older ages significantly increase the probability of late diagnosis [8]. However, the majority of studies concentrated in African or European countries, and there is no much information on the status of diagnosis of HIV in other regions and countries including Iran [9,10]. Iran, as one of the countries involving in HIV pandemics, is located in Western Asia. Based on the latest statistics in 2018, 61,000 Iranians are living with HIV. Moreover, this report adds that 2600 people died from AIDS-related diseases. Most cases occur in men, ages 15 and 44 years, and drug injection is the most common way for HIV transmission. Accordingly, Iran possesses a concentrated epidemic, which is the largest HIV epidemic among Middle East countries. Moreover, the report shows that 36% of people with HIV aware of their status and only 20% of the patients receive treatment and only 17% of people living with HIV were virally suppressed. Moreover, the statistic states that the majority of the patients (53%) had late HIV diagnosis [11]. However, statistical modeling suggests a number much more than the reported number for Iran, over 75,000 cases and even more [12]. However, due to the religious and cultural limitations in Iran, no comprehensive information is available about AIDS, and especially about the status of late diagnosis of HIV and the related-factors. In this study, we aimed to identify factors that contribute to the late diagnosis of HIV and survival of people with HIV in Iran.

Methods
The required data for this study was obtained from behavioral disease counseling centers (BDCC), which are established all over Iran. These centers responsible for collecting the demographic and clinical information include age, sex, the date of HIV diagnosis, the mode of HIV transmission, HIV status, the clinical stage of HIV/ AIDS, ART reception, CD4 cells count, education level, and marital status of the patients [13]. Inclusion criteria include all HIV-positive patients aged ≥16 years old, which have CD4 cell count within 3 months after diagnosis.
In this study, the late diagnosis was defined as a patient whose first CD4 cell count was less than CD4 < 350 cells/ mm3 within 90 days of diagnosis. Moreover, mortality data were extracted from BDCC. A subadministration in Behavioral disease counseling centers (BDCC) is responsible for registering of death events for each patient. Therefore, any death event is recorded by BDCC. To assure completeness of death registration, a double check is made with the Iranian Death Registration System (DRS), which is established to register death events in the general population and has appropriate coverage [13], and then the unmatched cases of death were added to death cases of BDCC.

Statistical analysis
First, we conducted a bivariate analysis to assess the crude association of risk factors with HIV late diagnosis and survival. In this step, any variable whose univariate test has a p-value < 0.25 was a candidate for the multivariate analysis. Hosmer and Lemeshow test was applied to check the goodness of fit of the final logistic model, which the model was found fit (p-value = 0.57). Also, Odds Ratios (OR) with 95% confidence intervals (95%CI) were used to summarize and describe the strength of association. Moreover, a Cox regression model was used to evaluate the factors affecting the survival of the patients. In the cox model, survival time (in months) was considered from the date of HIV diagnosis to the end of the follow-up period or the occurrence of death. Furthermore, for the patients who had not experienced death or lost to follow-up, the time from the date of HIV diagnosis to the end of the follow-up time was considered as the censored survival time. Moreover, Schoenfeld's residual test was used to investigate the proportional hazards (PH) assumption of the cox model.
We used the Amelia package in R4.0.3 to impute missing values. The percentage of missing values was 5% overall. Multiple imputations were applied to handle the missing data. In sensitivity analysis, the results of estimating factors associated with late HIV diagnosis and survival analysis were similar in both MI and complete case analyses. Data analyses were performed using Statistical Package for the Social Sciences (SPSS) version 22.0. In all analyses, a p-value less than 0.05 was considered statistically significant.

Descriptive and bivariate analysis
Data for 4402 patients between 1987 and 2016, whose first CD4 data were measured within 3 months after Based on the univariate analysis, late diagnosis was significantly higher in men, older age, people who inject the drug, people who have unprotected sexual contact, divorced persons, and the patients who have tuberculosis co-infection (p < 0.05). However, education level and job status were not significantly associated with late diagnosis (p > 0.05) ( Table 1).

Multivariate analysis Late HIV diagnosis
The result of the logistic regression model showed that the odds of late-diagnosis for men was 1.4 (95% CI: 1.15-1.67, p = 0.001) times that of women. In addition, the odds of late diagnosis in older people significantly were higher than younger people (Table 2).
TB co-infection was significantly associated with late diagnosis. People with TB significantly had higher odds for late HIV diagnosis than people without TB (OR = 2.06, 95%CI; 1.52-2.78).
Marital status and education level had not a significant association with late diagnosis (p > 0.05) ( Table 2).

Survival analysis
There was a statistically significant difference in the median survival time (day) between patients with early diagnosis (5009.53 ± 63.84) and late diagnosis (4404.90 ± 68.37). The Kaplan Meier plot for these two groups is shown in Fig. 1. Out of 674 patients who died, 492 (73%) had the late diagnosis, while 182 (27%) had the early diagnosis. Despite in total, 66.9% received ART, in subgroups, 31% of patients who died and 33.4% patients who survive received ART (Table 3).
In multivariate analysis, late diagnosis has a significant impact on the survival of patients. The odds of dying in people with late diagnosis were approximately two times higher than those with early diagnosis (Table 3). Furthermore, people aged 50 years and over are significantly less likely to survive than those aged under 30 years with HR = 1.81 (95% CI: 1.33-2.88). The odds of dying in men were 1.76 (95% CI: 1.28-2.42) times higher than women. The widows had higher odds for dying than the singles (1.68; 95% CI: 1.14-2.47). In addition, people with TB are at a higher risk of death compared to those without TB infection, with an HR = 1.77 (95% CI: 1.42-2.21). People diagnosed between 2006 and 2011 were significantly more at risk of death than those diagnosed after 2011 with a 2.45 (95% CI: 1.97-3.04) times higher odds of dying compared to after 2011. However, job status has no significant effect on the survival of patients (p > 0.05) ( Table 3).

Discussion
This is the first study investigating the status of late HIV diagnosis and its factors in Iran. We found that late diagnosis was prevalent among Iranian patients (53%). Existing evidence show that prevalence of late HIV diagnosis is high in other studies. Accordingly, the reported prevalence ranged from 25% in Columbia to 70.1% in China [14,15]. It is a major concern for the country because people with late HIV diagnosis receive ART late and, as a result, loses opportunities for counseling, education, and substance abuse treatment, resulting in the spread of HIV in the community. Evidence confirms that ART has the maximum effect on viral load and mortality rate from AIDS in the early stages of the disease. Therefore, ART is recommended to be initiated for all HIVpositive people immediately after diagnosis [16].
Moreover, we found that people who inject drugs (PWID) had the highest risk of late diagnosis. Like our study, a study conducted in China show that odds of late diagnosis in PWID was two times higher than other people [17]. Due to the severe social discrimination and stigma against both drug abuse and HIV in Iran, PWID avoid seeking medical services, including tests for HIV diagnosis. The literature has well established that ART is less prevalent in PWID than people who do not inject drugs. Additionally, in counties with free health services for HIV patients, medical care utilization is very low due to the patients' social situations and family problems [18,19].
In this study, the probability of late diagnosis was significantly greater in Iranian men than women. Other studies confirmed that male gender is a major risk factor of risky behaviors such as HIV late diagnosis [20].  that male gender is significantly associated with a higher risk of late HIV diagnosis [21][22][23]. This result could be attributed to two reasons. First, most PWID are men on the one hand, and late diagnosis is highly prevalent among PWID on the other hand. Therefore, men PWID are dominant in the AIDS population and have a higher frequency of late diagnosis than women [24]. Second, studies show that men generally are less likely to seek out health care than women. These studies have highlighted the fear of developing a disease as a significant barrier to seek medical care in men [25]. We found that the risk of late diagnosis was higher in older patients than in younger ones. A review of the literature also confirms that older age is the predominant predictor of late presentation of HIV [16,22,23,26]. Studies show that older patients have a low-risk perception relative to HIV, and therefore are less likely to be tested for HIV [8]. Late diagnosis is a significant factor of mortality in older people, and in fact, these groups obtain the most significant benefit from ART compared to other age groups [27]. Low education and knowledge and low risk perception about the disease have been mentioned as the main reasons for late diagnosis in older people. Low education is also associated with low socioeconomic status, affecting health and medical services [28,29].
Another identified risk factor of late HIV diagnosis was co-infection with TB. Gesesew et al. showed that HIV people with TB co-infection were about 2 times at risk of late HIV diagnosis [9].
This study showed that the prevalence of late HIV diagnosis was the lowest in mother-to-child transmission compared to the other transmission modes. Currently, a program called PMTCT (prevention of mother-to-child transmission) is being implemented [30]. All mothers are tested for HIV during pregnancy and are provided with HIV counseling. Preventive treatment should be given to the baby if the mother's test is positive. All infants born to HIV-positive mothers are also screened for HIV prophylaxis after birth and HIV. If a baby is infected with HIV, they will receive the same care and treatment services as other HIV-positive patients for the rest of their lives. This program aims to reduce the rate of mother-to-child transmission, one of the goals of the HIV prevention and control program recommended to countries by WHO and UNAIDS [31]. This study confirms the effectiveness of PMTCT on early HIV diagnosis. Therefore, it is recommended to improve the program and expand it to all regions of Iran.
This study's results provide implications for Iranian policymakers and health providers. First, the percentage of late diagnosis is high and require more attention.   [32]. Moreover, we recommend performing the mass screening of HIV and expanding counseling centers for high-risk groups, including FSWs and PWID. Furthermore, the increased coverage of ART is recommended to improve patient survival. We used registered data for this study as one of the study's limitations. Moreover, we did not measure some critical variables in late diagnosis and survival. Therefore, we could not assess their effects on late diagnosis or survival and could not control their confounding effects. Furthermore, the precision and validity of our retrospective study's data are questionable because the data were not collected for research objectives. Accordingly, the retrospective studies may potentially produce selection bias or information bias, and therefore, may distract the final results. Additionally, registered data lack data verification and do not have complete data/follow-up.

Conclusion
The prevalence of late HIV/AIDS diagnosis is very high among Iranians. For breaking the disease transmission chain, early HIV diagnosis through facilitating access to HIV testing for PWID by incorporating communitybased outreach, drug abuse treatment, and syringe exchange programs is recommended. Moreover, to improve survival of the patients, and decelerate the progression of the disease, supplying ART for all patients with AIDS is crucial.
Abbreviations CDC: Centers for Disease Control and Prevention; ART: Antiretroviral Therapy; HAART: Highly Active Antiviral Therapy; BDCC: Behavioral disease counseling centers; DRS: Death Registration System; PWID: People who inject drugs