Because of the availability of modern combined antiretroviral therapy (ART), people living with HIV (PLWH) have experienced a reduction in overall mortality and incidence of AIDS-defining conditions, improving survival and quality of life (1–3). The efficacy of ART is traditionally evaluated using two parameters: immunological recovery, namely the increase in CD4+ T-cell count (CD4), and the viral suppression, defined as an HIV RNA load of < 50 copies/ml in plasma. The extent of a CD4 increase, however, might have interpatient variations depending on age, ethnicity, test imprecision, intercurrent infection, and relative leukocyte count (4, 5), and the immunological response in the course of ART can be heterogeneous, based on individual, clinical, and genetic factors (6–8). Despite this variability, there is general agreement that CD4 count is the main marker of immunological staging (9). However, some PLWH fail to reconstitute CD4 count, even in the course of successful ART. These people are referred to as immunological non-responders (INRs). A consensus on the definition of INR has not been reached yet, owing to consistent interpatient variability of different factors, including age, HIV RNA load and CD4 count before starting ART, coinfections with HCV, HBV, and CMV, bone marrow and thymic dysfunction, genetic factors, and immune activation (10–12). INRs have been defined in the past based on CD4 count relative or absolute recovery. In studies that use relative CD4 count recovery from baseline, thresholds varied from less than 20% to 25% or 30% (13–15). Other studies consider the achievement of a predefined CD4 value, with thresholds that range from 200 to 500 cells/mm³ (16–18).

Each classification has pros and cons because of variable sensitivity and specificity. The choice of > 500 CD4+ T cells/mm³ to define full immune recovery could be justified in comparison to the general population, as it is the approximately the minimum expected CD4 count in a healthy, HIV-uninfected person. In addition, when counts are > 500 CD4+ T cells/mm³ in PLWH, their mortality rates become similar to those of the general population (19) and both mortality from AIDS- and non-AIDS-defining causes decreases as CD4 counts increase (20). On the other hand, the threshold of > 200 CD4+ T cells/mm³ might better represent the turning point between high and low risk of opportunistic diseases and AIDS-defining conditions (21). However, despite different definitions, most studies agree that INRs are at higher risk of disease progression toward AIDS-defining conditions, non-AIDS-defining conditions, and death (11, 14, 22). Because of the heterogenicity of definitions, the prevalence of INRs varies from 10% to 40% in different studies (13, 23); in the modern ART era INRs are expected to become rarer thanks to early diagnosis and immediate initiation of ART (24), and new antiretroviral agents characterized by high levels of efficacy and tolerability (25, 26).

In Italy, as well as in many other countries in the world, after 2008 new antiretroviral agents were introduced in clinical practice that completed the therapeutic armamentarium, and integrase strand transfer inhibitors (INSTIs) have begun to enter first-line therapy in PLWH (27, 28). With the aim of assessing the phenomenon of INR in an era of modern ART, we used three different definitions of immunological reconstitution, i.e., a CD4 level increase of > 20% from baseline, a CD4 level of > 200 cells/mm³ or a CD4 level of > 500 cells/mm³, and we evaluated the prevalence of INR according to these definitions in recent years in our single-center experience.

During the study period, 452 PLWH were newly diagnosed in our center. Among them, 128 were excluded from the study because they were lost to follow-up (N = 98), died (N = 20), or did not meet virological criteria for study inclusion (N = 10) within 36 months of starting ART.

The remaining 324 PLWH were included in the study. Among them, 91 out of 324 (28%) were female and 240 out of 324 (74%) were Italian. The median age of study participants was 41 years (IQR 18–80 years). The risk factors for HIV acquisition were as follows: unprotected intercourse, 278 PLWH (heterosexual, 183 [56%]; men who had sex with men [MSM], 95 [29%]); intravenous drug use (IVDU), 29 (9%) PLWH; and other/unknown, 17 (5%) PLWH.

The mean CD4 nadir of study participants before they received ART was 276 (± 222) cells/mm³, and the mean CD4 count after 36 months of ART was 664 (± 326) cells/mm³.

According to the CDC classification, 121 (37%), 131 (40%), and 71 (22%) PLWH were on stage A, B, and C, respectively, at time of diagnosis ( Table 1 ).

In the present study we found a prevalence of INR ranging between 1.5% and 34.6%, exploring different definition of immunological non-response.

The most restrictive definition that we used, namely that of INR₂₀₀, revealed an INR₂₀₀ rate of only 1.5%, which was lower than that reported in older studies that used the same threshold, which ranged between 15% and 26% (11, 42, 44), but which is consistent with the results of others (45). However, comparison among studies might be influenced by different study periods and antiretrovirals used in different settings and by heterogeneous duration of follow-up. The higher frequency of INR₂₀₀ found in previous years might reflect the use of older drugs that are currently considered sub-optimal or less tolerable, possibly compromising overall ART efficacy and, consequently, also immunological recovery (11, 42, 44). On the contrary, newer drugs such as INSTIs (46, 47), and also PIs, when compared with older ART approaches (48), might favor a more effective immune recovery. However, in our study, we did not find a consistent correlation between INR rate and months of exposure to different classes of antiretrovirals. It is notable that the study was conducted in an era of modern ART; if different ART classes were used in first-line therapy, the usefulness of comparing newer and older drugs is limited by the absence of an historical control group. In addition, the observation about the role of a low CD4 nadir underlying the INR₂₀₀ phenomenon underscores the very important role of early diagnosis in preventing the INR phenomenon; increased awareness of the risk of infection and improved timing of diagnosis could explain, at least in part, the reduction in the current number of INR₂₀₀.

On the other hand, when assessing frequency of INR₅₀₀, we still found about 35% of PLWH with a discordant viro-immunological response, in line with the frequency found in previous studies (34, 45, 46), suggesting that, even if modern therapies have contributed to make INR₂₀₀ rarer, the complete viro-immunological response is still not guaranteed nor consistently improved compared with the past, when using the CD4 threshold of 500 cells/mm³. Even if the highest risk of clinical events and death has been associated with INR₂₀₀ (11, 42, 44), PLWH with a CD4 count in the range 200–499 cells/mm³ still have higher mortality rates than those with a CD4 count > 500 cells/mm³ (49). Therefore, the goal of ART should be to get patients over this threshold and reaching an immunological condition as close as possible to that of the general population. Moreover, INR₅₀₀ have been shown to have higher percentages of activated CD4+ T cells, regulatory T cells (T_reg), effector T_reg and terminal effector T_reg suggesting, in this group of PLWH, a residual immune activation persisting after many years of ART (18, 35). In addition, immune responses against vaccines are still different among PWH with CD4+ T cell counts of < 200, 200–500, or > 500 cells/mm³, and in those with counts of > 500 cells/mm³ responses comparable to those in the HIV-uninfected population have been described (50). The persistence of a certain grade of inflammation might also be indirectly inferred from the CD4/CD8 ratio, a surrogate marker of T-cell compartment balance, reflecting both CD4 T-cell recovery and CD8 T-cell activation, expansion, and senescence (51, 8). Few studies have investigated the relationship between CD4/CD8 ratio and immune recovery, supporting an association between lower baseline levels and immunodiscordant response to ART (46, 52). In our study, we found not only that a higher CD4/CD8 ratio was protective toward becoming either INR₂₀₀ or INR₅₀₀ at time of ART initiation, but also that that CD4/CD8 ratio was more likely to be lower after 36 months of ART in INR₂₀₀ and INR₅₀₀ than in full responders, supporting the hypothesis of a possible persistent immune activation impairing reconstitution of the immune system and CD4 gains (53). Moreover, CD4 nadir was confirmed to be the stronger predictor of immunological non-response in INR₂₀₀ and INR₅₀₀, in accordance with data from the literature (18, 34, 44, 45, 54), and all other epidemiological and clinical variables considered did not consistently correlate with INR. Instead, these data were not confirmed when using the definition of INR_20%. In fact, in our study, the odds of being INR_20% was higher in PLWH with a higher CD4 nadir and higher CD4/CD8 ratio. Being INR_20% has been described in the past as a predictor of clinical progression (14); however, the definition of INR_20% is highly influenced by baseline CD4– T cell count and, in an era of modern ART, where therapy is initiated regardless of total CD4 count at HIV diagnosis, this definition may be misleading, as it may imply the inclusion of early-treated patients with unimpaired baseline CD4 counts. Consequently, these PLWH do not experience an increase in CD4 above the predefined threshold of 20% from baseline, but this should not be interpreted as a poor response. On the contrary, for PLWH with extremely low CD4 counts, a small CD4 gain could even exceed the threshold of 20% while remaining with a very low absolute number of lymphocytes and CD4/CD8 ratio. Therefore, even if the definition can still be applied to certain clinical situations in PLWH stratified based on baseline CD4 counts, or in certain research contexts, INR_20% might not be a suitable definition for identifying in general PLWH with poor responses in clinical practice.

The present study has several limitations. The retrospective design and relatively small sample size of the cohort limit the strength and generalizability of the findings. Moreover, data on biomarkers of immune activation, on viral reservoir, as well as data on herpetic viral co-infections, genetics, and behavioral or dietary factors were not available, limiting the possibility of investigation of further variables influencing immunological responses. In addition, the exclusion of people lost to follow-up or dead before the predefined time point of 36 months might have contributed to underestimation of INRs. Despite these limits, the study highlights how INR₂₀₀ have become very rare in the contemporary ART era, and still about one-third of PLWH meet the criteria for INR₅₀₀. Overcoming the threshold of 500 CD4/mm³ could be more appropriate to define full responders, in contrast with the older definitions of INR₂₀₀ and INR_20%. Although our results do not show a benefit from choosing different ART strategies to improve immune recovery, early diagnosis, and rapid treatment initiation before CD4 counts and CD4/CD8 ratio begin to decline are critical to achieving an optimal immune response.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ligurian Ethics Committee. The patients/participants provided their written informed consent to participate in this study.

LT, LL, FB, GB, SB, AB, and MB performed the research. SM, SB, and MG managed the database and checked the accuracy of the data. LT, LL, and AB designed the research study. LT analyzed the data. LT, LL, FB, and GB wrote the paper. MB and AB reviewed the final version of the paper and the scientific contents of the study. All authors have read and approved the final version of the paper.