CD4/CD8 Ratio and the Risk of Kaposi Sarcoma or Non-Hodgkin Lymphoma in the Context of Efficiently Treated Human Immunodeficiency Virus (HIV) Infection: A Collaborative Analysis of 20 European Cohort Studies

Abstract

Background

A persistently low CD4/CD8 ratio has been reported to inversely correlate with the risk of non-AIDS defining cancer in people living with human immunodeficiency virus (HIV; PLWH) efficiently treated by combination antiretroviral therapy (cART). We evaluated the impact of the CD4/CD8 ratio on the risk of Kaposi sarcoma (KS) or non-Hodgkin lymphoma (NHL), still among the most frequent cancers in treated PLWH.

Methods

PLWH from the Collaboration of Observational HIV Epidemiological Research Europe (COHERE) were included if they achieved virological control (viral load ≤ 500 copies/mL) within 9 months following cART and without previous KS/LNH diagnosis. Cox models were used to identify factors associated with KS or NHL risk, in all participants and those with CD4 ≥ 500/mm³ at virological control. We analyzed the CD4/CD8 ratio, CD4 count and CD8 count as time-dependent variables, using spline transformations.

Results

We included 56 708 PLWH, enrolled between 2000 and 2014. At virological control, the median (interquartile range [IQR]) CD4 count, CD8 count, and CD4/CD8 ratio were 414 (296–552)/mm³, 936 (670–1304)/mm³, and 0.43 (0.28–0.65), respectively. Overall, 221 KS and 187 NHL were diagnosed 9 (2–37) and 18 (7–42) months after virological control. Low CD4/CD8 ratios were associated with KS risk (hazard ratio [HR] = 2.02 [95% confidence interval {CI } = 1.23–3.31]) when comparing CD4/CD8 = 0.3 to CD4/CD8 = 1) but not with NHL risk. High CD8 counts were associated with higher NHL risk (HR = 3.14 [95% CI = 1.58–6.22]) when comparing CD8 = 3000/mm³ to CD8 = 1000/mm³). Similar results with increased associations were found in PLWH with CD4 ≥ 500/mm³ at virological control (HR = 3.27 [95% CI = 1.60–6.56] for KS; HR = 5.28 [95% CI = 2.17–12.83] for NHL).

Conclusions

Low CD4/CD8 ratios and high CD8 counts despite effective cART were associated with increased KS/NHL risks respectively, especially when CD4 ≥ 500/mm³.

Kaposi sarcoma, non-Hodgkin lymphoma, CD4/CD8 ratio, CD8 T-cells, efficient cART

Despite the widespread use of effective combination antiretroviral therapy (cART), a higher incidence of AIDS- and non-AIDS-defining events persist in people living with human immunodeficiency virus (HIV; PLWH), compared to the general population [1–3]. These appear to be driven by persistent immune activation, including inflammation, despite long-term viral suppression and CD4 restoration thanks to effective cART [4–6].

The CD4/CD8 ratio is considered as a reliable marker of systemic immune activation during successful cART [7, 8]. Low CD4/CD8 ratios are associated with higher CD4 and CD8 T-cell activation, exhaustion, and senescence, as well as with increased innate immune activation such as monocyte activation, even in individuals with undetectable viremia and CD4 ≥ 500/mm³ [9, 10]. Furthermore, in PLWH receiving efficient cART, the CD4/CD8 ratio inversely correlates with the risk of AIDS-related mortality, and the risk of non-AIDS-defining events, especially non-AIDS-defining cancers [11–14]. However, the impact of the CD4/CD8 ratio on the risk of AIDS-defining cancers, such as Kaposi sarcoma (KS) and non-Hodgkin lymphoma (NHL), has never been studied.

KS and NHL remain among the most frequent cancers in PLWH receiving cART [15, 16]. KS is an angioproliferative malignancy associated with human herpesvirus-8 (HHV-8), whereas NHL can be induced by either Epstein-Barr virus (EBV), HHV-8, or human T-cell leukemia/lymphoma virus type 1 (HTLV-1), or it may not be viro-induced, depending on subtypes. A previous cohort study conducted in the French Hospital Database on HIV (FHDH-ANRS CO4) found a 35-fold increased KS risk in PLWH relative to that of the general population, although viremia was suppressed and CD4 counts restored to >500/mm³ for ≥ 2 years. For NHL, the risk was 9-fold higher in PLWH than the general population but dropped to that observed in the general population when including only PLWH with suppressed viremia and a stably restored CD4 count [3].

Using data from a large collaboration of European cohort studies of PLWH (COHERE [Collaboration of Observational HIV Epidemiological Research Europe]), we aimed to evaluate the impact of CD4/CD8 ratio restoration on KS or NHL risk in PLWH on efficient cART, particularly whether the CD4/CD8 ratio provides additional information to the CD4 count, which is the usual immunological predictor for KS and NHL. As initiating cART with high CD4 count is becoming increasingly frequent in the era of early cART recommendations, a secondary analysis was conducted in PLWH with CD4 ≥ 500/mm³ at virological control [17].

METHODS

The COHERE in EuroCoord (www.cohere.org) is a European collaboration of cohorts of people living with HIV (PLWH) in routine clinical care. A detailed description of COHERE has been previously published [18, 19]. For the current project, data were merged in 2015 in COHERE in EuroCoord, and 27 European cohorts provided data from 149 028 PLWH aged ≥16 years.

INCLUSION CRITERIA AND DEFINITIONS

Human immunodeficiency virus type 1 (HIV-1) infected adults who were ≥16 years old, initiated cART between 1 January 2000 (when cART became available in all parts of Europe) and 31 December 2014, and achieved virological control within 9 months following cART initiation were included. Virological control was defined as the first plasma HIV-1 RNA ≤ 500 copies/mL. We used a threshold at 500 copies/mL to account for the heterogeneity of the limit of detection used in methods measuring plasma HIV-1 RNA during the study period and between participating cohorts. At least one CD4/CD8 measurement was required within 6 months after virological control. Cohorts were eligible if CD8 count measurements were reported for ≥70% of blood samples used to measure the CD4 count, to avoid the selection of populations where CD8 measurements would have been performed for medical reason, which could have led to a selection bias. Overall, 20 of the 27 participating cohorts COHERE were selected.

For both the main and secondary analyses, the baseline was defined as the time of the first CD4/CD8 ratio measurement within 6 months after virological control. In the secondary analysis, the CD4 count of the first ratio measurement had to be ≥500/mm³. PLWH with a diagnosis of KS or NHL at baseline or before were excluded.

Statistical Analyses

Continuous variables are expressed as medians and interquartile ranges (IQR) and categorical variables as counts and percentages.

We used separate Cox analyses stratified by cohort, to identify factors associated with KS or NHL risk in the study population. Follow-up was censored at the date of patient’s last clinic visit.

Sociodemographic characteristics, baseline characteristics, and immunovirological factors were included in univariable analyses. (i) Sociodemographic characteristics were age at baseline and a composite variable combining sex, HIV exposure risk, and geographical origin, consisting of 6 categories: men who have sex with men (MSM) regardless of geographical origin, intravenous drug users (IDU) regardless of geographical origin, sub-Saharan women, heterosexual sub-Saharan men, non-sub-Saharan women, and heterosexual non-sub-Saharan men. (ii) Baseline characteristics were clinical Center for Disease Control and Prevention (CDC) stage C, hepatitis B virus (HBV) coinfection (defined by positive hepatitis B surface [HBs] antigen), hepatitis C virus (HCV) coinfection (defined by positive anti-HCV immunoglobulin G [IgG] antibodies), and calendar period of cART introduction (2000–2004; 2005–2009; 2010–2014). (iii) Immunovirological factors were CD4 count, CD8 count, CD4/CD8 ratio, and virological failure. These factors were considered as time-dependent variables for the analyses. To that end, time was divided into successive 6-month periods during the first 3 years of follow-up and 12-month periods thereafter. Indeed, after 3 years of follow-up, the time between visits was >6 months for 25% of PLWH. For each period, we used the first value of the CD4/CD8 ratio, corresponding CD4 and CD8 counts, and maximal plasma HIV-1 RNA value. Virological failure was defined as plasma HIV-1 RNA > 500 copies/mL.

We analyzed the CD4/CD8 ratio, CD4 count, and CD8 count as continuous variables, using spline transformations, in order to capture nonlinear associations with KS or NHL risk. The knots for the restricted cubic splines were positioned by default, using Harrell recommended percentiles [20, 21]. We used tertiles of 0.43, 0.64, 0.91 for the CD4/CD8 ratio; 404, 536, 700/mm³ for the CD4 count; and 621, 854, 1165/mm³ for the CD8 count, in the main analysis. For the secondary analysis, tertiles were 0.55, 0.79, 1.10 for the CD4/CD8 ratio;, 568, 704, 881/mm³ for the CD4 count; and 669, 916, 1246/mm³ for the CD8 count. Other knots based on clinical cutoffs were also tested and gave similar results.

Three models were tested to better understand the impact of immune factors on KS and NHL risk. The first (MV1) was adjusted for the CD4/CD8 ratio, the second (MV2) for both CD4 count and CD8 count, and the third (MV3) for both the CD4/CD8 ratio and CD4 count. All 3 models were additionally adjusted for variables with a univariable P value < .20, namely, age at baseline, sex, group of HIV exposure risk, origin, calendar period of cART introduction, and virological failure. A P-value < .05 was considered significant. Akaike’s information criterion (AIC) was used to compare the fit of the 3 statistical models: the lower the AIC value, the better the fit. A secondary analysis was conducted in PLWH with CD4 ≥ 500/mm³ at baseline. To test the robustness of our results in PLWH without virological failure, post hoc analyses were performed with censored follow-up not only at patient’s last clinic visit but also at virological failure. Other post hoc analyses restricted to patients starting cART from 2005, where the limit of HIV-RNA assay was 50 copies/mL or lower in 99% of measurements, were carried out to evaluate the impact of low viremia (50–500 copies/mL) on KS/NHL risks.

SAS statistical software version 9.4 (SAS Institute, Inc, Cary, North Carolina, USA) was used for all analyses.

RESULTS

Study Population

Overall, 56 708 PLWH from 20 cohorts in 12 countries (France, Austria, Greece, the Netherlands, Germany, Italy, Spain, Denmark, Switzerland, Belgium, Sweden, and United Kingdom) were eligible for the current analyses. Figure 1 shows the study flowchart. The main reasons for excluding patients were being enrolled in cohorts where CD8 count measurements were recorded in <70% of samples used for CD4 count measurements (n = 21 024), having started cART before 2000 (n = 36 126), and achieving virological control after 9 months following cART introduction (n = 21 127). In addition, 1600 PLWH with a KS or NHL diagnosis at baseline or before were excluded.

Figure 1.

Study flow-chart. Abbreviations: cART, combination antiretroviral therapy; COHERE, Collaboration of Observational HIV Epidemiological Research Europe; HIV-1, human immunodeficiency virus type 1; KS, Kaposi sarcoma; NHL, non-Hodgkin lymphoma.

Baseline characteristics are shown in Table 1. Overall, 27 540 (49%) were MSM, and 14 259 (25%) were women. Most PLWH were of European origin or from other western countries (including Australia and New Zealand) (n = 39 002; 69%), whereas 8461 (15%) were of sub-Saharan origin. The median time from cART introduction to baseline was 2.5 months (IQR = 1.2–4.0). At baseline, participants had a median CD4 count of 414/mm³ (IQR = 296–552), a median CD8 count of 936/mm³ (IQR = 670–1304), and a median CD4/CD8 ratio of 0.43 (IQR = 0.28–0.65). Only 8% of the study population had a CD4/CD8 ratio ≥ 1, whereas 59% had a very low CD4/CD8 ratio (<0.5) at baseline, and 19 133 (34%) PLWH had a CD4 count of ≥500/mm³.

Table 1.

Characteristics of the Study Population at Baseline^a (n = 56 708 PLWH)

Characteristicsb
Age (years)	38 (32–45)
Sex, HIV exposure risk, geographical origin
MSM	27 540 (49)
Intravenous drug users	3015 (5)
Sub-Saharan women	5770 (10)
Non-sub-Saharan women	7690 (13)
Heterosexual sub-Saharan men	2691 (5)
Heterosexual non-sub-Saharan men	10 002 (18)
HBs antigen (n missing = 25 000)
Negative	30 230 (95)
Positive	1478 (5)
Anti-HCV IgG (n missing = 25 397)
Negative	27 285 (87)
Positive	4026 (13)
Clinical CDC stage C	7776 (14)
CD4 count (/mm3)
<200	6133 (11)
200–350	14 073 (25)
350–500	17 369 (30)
500–650	10 782 (19)
≥650	8351 (15)
CD8 count (/mm3)
<1000	31 354 (55)
1000–2000	21 844 (39)
≥2000	3510 (6)
CD4/CD8 ratio
c<0.3	16 036 (28)
.3–.5	17 759 (31)
.5–.8	14 388 (26)
.8–1.0	3980 (7)
≥1.0	4545 (8)
Calendar period of cART introduction
2000–2004	16 924 (30)
2005–2009	24 333 (43)
2010 or later	15 451 (27)

Abbreviations: cART, combination antiretroviral therapy; CDC, Centers for Disease Control and Prevention; HBs antigen, hepatitis B surface antigen; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IgG, immunoglobulin G; MSM, men who have sex with men; PLWH, people living with human immunodeficiency virus.

^aBaseline is the date of the first CD4/CD8 measurement within 6 months following virological control.

^bResults are expressed as median (interquartile range) or n (%).

Overall, participants were followed up for a median of 59 (IQR = 30–96) months from baseline, accounting for 307 700 person-years (py). The median number of CD4 count, CD8 count, and CD4/CD8 ratio measurements per patient were 15 (8–25), 14 (7–24), and 14 (7–24), respectively. The cumulative incidence of CD4/CD8 ratio restoration ≥1 from baseline was 28% (95% confidence interval [CI] 27–28) at 2 years and 45% (95% CI 44–46) at 5 years. The corresponding cumulative incidence among PLWH with CD4 ≥ 500/mm³ at baseline, was 46% (95% CI 45–47) and 63% (95% CI 62–64), respectively.

A total of 221 KS and 187 NHL were diagnosed after baseline, corresponding to incidence rates of 72/100 000 py (95% CI 55–89) for KS and 61/100 000 py (95% CI 46–76) for NHL. KS were diagnosed 9 (IQR = 2–37) months after baseline, whereas NHL diagnoses occurred 18 (IQR = 7–42) months after baseline. KS and NHL occurred in the context of virological failure for 33/221 (15%) and 23/187 (12%) PLWH, respectively.

The secondary analysis included 19 133 PLWH-1 with CD4 ≥ 500/mm³ at baseline, of whom 65 presented with KS and 50 with NHL. KS and NHL occurred in the context of virological failure for 14 (21%) and 11 (22%) PLWH, respectively.

Factors Associated With the Risk of Kaposi Sarcoma (Table 2)

Immunovirological factors associated with KS risk in the 3 different multivariable models MV1, MV2, and MV3 are presented in Table 2. Virological failure was the strongest factor associated with KS risk in all 3 models, with a nearly 3-fold higher risk.

Table 2.

Immune-Virological Factors Associated With the Risk of Kaposi Sarcoma: Multivariable Cox Analyses

	Primary Analysis in the Whole Study Population (n = 56 708 PLWH, n = 221 incident KS)			Secondary Analysis in PLWH With CD4 ≥ 500/mm3 at Baselinea (n = 19 133 PLWH, n = 65 incident KS)
	MV1b AIC = 3419	MV2b AIC = 3402	MV3b AIC = 3393	MV3b
	Adjusted Hazard Ratio (95% CI)	Adjusted Hazard Ratio (95% CI)	Adjusted Hazard Ratio (95% CI)	Adjusted Hazard Ratio (95% CI)
CD4/CD8 ratioc from baselinea
CD4/CD8 = 1.5	.72 (.51–1.02)	Not included in the model	.80 (.56–1.10)	.60 (.31–1.16)
CD4/CD8 = 1.2	.85 (.73–.99)		.88 (.76–1.03)	.78 (.58–1.05)
CD4/CD8 = 1	1		1	1
CD4/CD8 = .8	1.25 (1.05–1.49)		1.18 (.98–1.42)	1.38 (1.00–1.91)
CD4/CD8 = .5	1.87 (1.28–2.69)		1.64 (1.10–2.45)	2.35 (1.33–4.14)
CD4/CD8 = .3	2.56 (1.68–3.92)		2.02 (1.23–3.31)	3.27 (1.60–6.56)
CD4 countc (/mm3) from baselinea
CD4 = 350/mm3	Not included in the model	1.78 (1.25–2.53)	1.57 (1.09–2.25)	1.81 (1.04–3.14)
CD4 = 500/mm3		1	1	1
CD4 = 650/mm3		.63 (.49–.82)	.71 (.54–.94)	.77 (.44–1.36)
CD4 = 800/mm3		.54 (.32–.93)	.68 (.39–1.20)	.49 (.22–1.09)
CD4 = 1000/mm3		.54 (.28–1.01)	.74 (.38–1.45)	.33 (.11–.97)
CD8 countc (/mm3) from baselinea
CD8 = 300/mm3	Not included in the model	1.01 (.76–1.35)	Not included in the model	Not included in the model
CD8 = 500/mm3		1
CD8 = 800/mm3		1.22 (.89–1.66)
CD8 = 1000/mm3		1.33 (.97–1.83)
CD8 = 1200/mm3		1.43 (1.02–1.99)
CD8 = 1600/mm3		1.58 (1.07–2.32)
CD8 = 2000/mm3		1.67 (1.10–2.53)
Virological failured from baselinea	3.11 (2.13–4.52)	2.88 (1.98–4.19)	2.77 (1.90–4.06)	2.69 (1.31–5.29)

Abbreviations: AIC, Akaike information criterion; CI, confidence interval; KS, Kaposi sarcoma; MV1, multivariable analysis 1; MV2, multivariable analysis 2; MV3, multivariable analysis 3; PLWH, people living with human immunodeficiency virus.

^aBaseline is the date of the first CD4/CD8 measurement within 6 months following virological control.

^bMV1 was adjusted for the CD4/CD8 ratio, MV2 was adjusted for both CD4 and CD8 counts, and MV3 was adjusted for both the CD4/CD8 ratio and CD4 count. All 3 models were additionally adjusted for variables with a univariable P value < .20: age at baseline, sex, group of human immunodeficiency virus (HIV) exposure risk, origin, calendar period of combination antiretroviral therapy (cART) introduction, and virological failure^d. Virological failure^d, CD4 count, CD8 count, and CD4/CD8 ratio were considered as time-dependent variables from baseline. All analyses were stratified by cohort.

^cCD4 count, CD8 count, and CD4/CD8 ratio were modeled using spline transformations.

^dVirological failure is defined as plasma human immunodeficiency virus type 1(HIV-1) RNA > 500 copies/mL after baseline^a.

In MV1, adjusted for CD4/CD8 ratio as the only immune factor, the lower the CD4/CD8 ratio the higher the KS risk (hazard ratio [HR] from 1.25 for PLWH-1 with CD4/CD8 = 0.8, to 2.56 for those with CD4/CD8 = 0.3, when compared to CD4/CD8 = 1).

In MV2, which included the CD4 and CD8 counts separately, without the CD4/CD8 ratio, both immune parameters (low CD4 counts and high CD8 counts) were independently associated with KS risk.

MV3 was adjusted for both CD4 count and CD4/CD8 ratio and confirmed independent associations of both parameters with KS risk: CD4 restoration was associated with decreased KS risk, even for CD4 > 500/mm³ (HR = 1.57 when CD4 = 350/mm³ to HR = 0.71 when CD4 = 650/mm³, when comparing to CD4 = 500/mm³). Furthermore, the CD4/CD8 ratio was gradually associated with KS risk (HR from 1.18 in PLWH with CD4/CD8 = 0.8, to 2.02 for those with CD4/CD8 = 0.3, when comparing to CD4/CD8 = 1). According to the AIC value, MV3 was the most accurate model (Table 2).

Other independent factors associated with higher KS risk were being MSM and being older (Supplementary Table 1).

In PLWH with CD4 ≥ 500/mm³ at baseline, the inverse association between CD4/CD8 ratio and KS risk was stronger than that observed in the entire study population. This is illustrated by Figure 2A (overall analysis) and Figure 2B (analysis restricted to PLWH with CD4 ≥ 500/mm³ at baseline) showing the impact of CD4/CD8 ratio on KS risk according to the most accurate model adjusted for both CD4 and CD4/CD8 ratio (MV3).

Figure 2.

Effect of the CD4/CD8 ratio restoration on KS risk, in the whole study population (A) and in PLWH with CD4 ≥ 500/mm³ at baseline* (B). Abbreviations: cART, combination antiretroviral therapy; CI, confidence interval; HR, hazard ratio; KS, Kaposi sarcoma; PLWH, people living with human immunodeficiency virus. *Baseline is the date of the first CD4/CD8 measurement within six months following virological control. **Plots of computed HR from multivariable Cox analysis adjusted for age, sex, group of human immunodeficiency virus (HIV) exposure risk, origin, calendar period of cART introduction, virological failure, CD4 count (multivariable analysis 3 [MV3]). Virological failure, CD4 count, and CD4/CD8 ratio were considered as time-dependent variables from baseline. CD4 count and CD4/CD8 ratio were modeled using spline transformations.

Factors Associated With the Risk of Non-Hodgkin Lymphoma (Table 3)

As observed with KS, virological failure was strongly associated with NHL risk, with a 2-fold higher risk in all multivariable models.

Contrary to KS, the CD4/CD8 ratio did not add supplementary predictive information to the CD4 count for NHL risk. However, MV2 showed that NHL risk strongly increased when the absolute CD8 count was very high (HR from 1.61 for CD8 = 2000/mm³, to 3.14 for CD8 = 3000/mm³, compared to CD8 = 1000/mm³). According to the AIC, MV2, adjusted for both CD4 count and CD8 count, was the most accurate model (Table 3).

Table 3.

Immune-Virologocal Factors Associated With the Risk of Non-Hodgkin lymphoma: Multivariable Cox Analyses

	Primary Analysis in the Whole Study population (n = 56 708 PLWH, n = 187 incident NHL)			Secondary Analysis in PLWH With CD4 ≥ 500/mm3 at Baselinea (n = 19 133 PLWH, n = 50 incident NHL)
	MV1b AIC = 2819	MV2b AIC = 2770	MV3b AIC = 2772	MV2b
	Adjusted HR (95% CI)	Adjusted HR (95% CI)	Adjusted HR (95% CI)	Adjusted HR (95% CI)
CD4/CD8 ratioc from baselinea
CD4/CD8 = 1.5	.92 (.69–1.21)	Not included in the model	.90 (.66–1.23)	Not included in the model
CD4/CD8 = 1.2	.95 (.84–1.08)		.96 (.83–1.11)
CD4/CD8 = 1	1		1
CD4/CD8 = .8	1 .07 (.91–1.27)		1.04 (.87–1.24)
CD4/CD8 = .5	1.41 (.97–2.04)		1.24 (.84–1.83)
CD4/CD8 = .3	2.25 (1.45–3.49)		1.53 (.93–2.52)
CD4 countc (/mm3) from baselinea
CD4 = 350/mm3	Not included in the model	1.26 (.87–1.84)	1.21 (.83–1.76)	3.15 (1.53–6.52)
CD4 = 500/mm3		1	1	1
CD4 = 650/mm3		.66 (.47–.94)	.70 (.49–1.00)	.47 (.23–.94)
CD4 = 800/mm3		.49 (.25–.95)	.56 (.28–1.11)	.38 (.15–.96)
CD4 = 1000/mm3		.76 (.42–1.38)	.92 (.47–1.79)	.50 (.21–1.21)
CD8 countc (/mm3) from baselinea
CD8 = 500/mm3	Not included in the model	1.05 (.75–1.45)	Not included in the model	1.02 (.45–2.32)
CD8 = 1000/mm3		1		1
CD8 = 1500/mm3		1.19 (.90–1.56)		1.36 (.79–2.34)
CD8 = 2000/mm3		1.61 (1.09–2.37)		2.29 (1.09–4.80)
CD8 = 2500/mm3		2.28 (1.38–3.78)		3.36 (1.55–7.29)
CD8 = 3000/mm3		3.14 (1.58–6.22)		5.28 (2.17–12.83)
Virological failured from baselinea	2.23 (1.39–3.58)	1.99 (1.22–3.23)	1.99 (1.22–3.25)	1.94 (.90–4.17)

Abbreviations: AIC, Akaike information criterion; CI, confidence interval; HR, hazard ratio; MV1, multivariable analysis 1; MV2, multivariable analysis 2; MV3, multivariable analysis 3; NHL, non-Hodgkin lymphoma; PLWH, people living with human immunodeficiency virus.

^aBaseline is the date of the first CD4/CD8 measurement within 6 months following virological control.

^bMV1 was adjusted for the CD4/CD8 ratio, MV2 for both CD4 and CD8 counts, and MV3 for both the CD4/CD8 ratio and CD4 count. All 3 models were additionally adjusted for variables with a univariable P value < .20: age at baseline, sex, group of human immunodeficiency virus (HIV) exposure risk, origin, calendar period of combination antiretroviral therapy (cART) introduction, and virological failure^d. Virological failure^d, CD4 count, CD8 count, and CD4/CD8 ratio were considered as time-dependent variables from baseline. All analyses were stratified by cohort.

^cCD4 count, CD8 count, and CD4/CD8 ratio were modeled using spline transformations.

^dVirological failure is defined as plasma human immunodeficiency virus type 1 (HIV-1) RNA > 500 copies/mL after baseline^a.

Other independent factors associated with higher NHL risk were being older and male sex (Supplementary Table 1).

In PLWH with CD4 ≥ 500/mm³ at baseline, the positive association between CD8 count and NHL risk was stronger than that observed for the entire study population. This is illustrated by Figure 3A (overall analysis) and Figure 3B (analysis restricted to PLWH with CD4 ≥ 500/mm³ at baseline), showing the impact of CD8 count on NHL risk, according to the most accurate model adjusted for both CD4 count and CD8 count (MV2).

Figure 3.

Effect of CD8 counts on NHL risk, in the whole study population (A) and in PLWH with CD4 ≥ 500/mm³ at baseline* (B). Abbreviations: cART, combination antiretroviral therapy; CI, confidence interval; HR, hazard ratio; NHL, non-Hodgkin lymphoma; PLWH, people living with human immunodeficiency virus. *Baseline is the date of the first CD4/CD8 measurement within 6 months following virological control. **Plots of computed HR from multivariable Cox analysis adjusted for age, sex, group of human immunodeficiency virus (HIV) exposure risk, origin, calendar period of cART introduction, virological failure, CD4 count (multivariable analysis 2 [MV2]). Virological failure, CD4 count, and CD8 count were considered as time-dependent variables from baseline. CD4 count and CD8 count were modeled using spline transformations.

Post hoc Analyses

When the follow-up of PLWH was censored at virological failure, similar associations were found between the CD4/CD8 ratio and KS risk, and between CD8 count and NHL risk (Supplementary Table 2). Analyses restricted to PLWH starting cART from 2005 showed no impact of low viremia (50–500 copies/mL) on KS/NHL risks (Supplementary Table 3).

DISCUSSION

In this large European clinical cohort collaboration, we observed a differential impact of the CD4/CD8 ratio on KS or NHL risk, both AIDS-defining cancers. A low CD4/CD8 ratio was associated with KS risk, even after adjustment for CD4 count and viral load. In contrast, the CD4/CD8 ratio was not associated with NHL risk after adjustment for CD4 count, whereas very high CD8 counts (≥2000/mm³) were strongly associated with NHL risk. As for other HIV-related morbidities, virological failure was strongly associated with both KS and NHL risks [3, 13].

Secondary analyses conducted in the subgroup of PLWH with CD4 ≥ 500/mm³ at baseline showed a stronger association of both the CD4/CD8 ratio for KS risk and the CD8 count for NHL risk.

Strength and Limitations

This is the first study to evaluate the impact of the CD4/CD8 ratio on the risk of AIDS-defining cancers. This analysis, based on a large cohort collaboration with a long-term follow-up, allowed the study of rare events separately, in the setting of efficient cART and CD4 ≥ 500/mm³. Our findings confirmed the importance of studying cancers separately, because the association between CD4/CD8 ratio and cancer risk differed between the 2. Limitations included the absence of information on KS clinical presentation, NHL histological subtypes, and viral serostatus associated with NHL (ie, EBV, HHV-8, HTLV-1), precluding analyses according to these parameters. Evaluating the association between CD4/CD8 ratio and the risk of immune reconstitution inflammatory syndrome was limited because baseline started close to 3 months after cART introduction, and most unmasking IRIS events occur within this time frame.

Kaposi Sarcoma

Despite large cART-related declines in KS incidence over time, a higher KS risk persists in PLWH than in the general population, especially for MSM and PLWH from sub-Saharan Africa, among whom HHV-8 seroprevalence is elevated [3, 15, 22, 23]. Nevertheless, the clinical context in which KS occurs has considerably changed during the effective cART era. A growing proportion of KS is diagnosed among patients on cART, with higher CD4 counts and/or suppressed viremia, showing different clinical presentations [24].

A previous study of effectively treated PLWH showed KS to be associated with an increased frequency of T cells with immunosenescent phenotypes and lower frequencies of naive T cells [25]. The mechanisms linking immune activation to KS include pro-inflammatory cytokines present in KS microenvironment and involved in the stimulation of angiogenic KS lesions. The KS microenvironment is characterized by inflammatory cellular infiltrates and low T-cell infiltrates, and is associated with very low anti-HHV8 CD8 T-cell responses in peripheral blood [26, 27]. Although the link between immune activation and KS development might be bidirectional, our findings suggest that HIV-related immune activation, reflected by CD4/CD8 ratio, may contribute to KS development in PLWH on efficient cART.

Non-Hodgkin Lymphoma

Efficient cART have reduced the risk of NHL and induced substantial shift in the subtypes of lymphoma observed in PLWH [24, 28]. Higher risks of NHL in PLWH on cART may be due to persistent low-level HIV replication on cART, resulting in chronic B-cell activation and promoting oncogenic mutations and translocations, leading to lymphomagenesis [29, 30]. Moreover, persistence of an HIV latent reservoir may also influence NHL risk, the HIV proteins promoting B-cell clonogenicity [31].

Here the CD4/CD8 ratio did not add supplementary predictive information to the CD4 count for NHL risk. Nevertheless, the highest CD8 counts (≥2000/mm³) were associated with elevated NHL risk. Only one nested case-control study has evaluated CD8 counts before NHL diagnosis in PLWH on cART, in which there was a greater decline in CD8 counts during the last 2 years preceding NHL diagnosis than that observed in controls (−184/mm³ vs −19/mm³) [32]. Absolute CD8 counts were not reported.

Other studies reported elevated levels of immune activation markers several years before NHL diagnosis, suggesting that immune activation likely plays a key role in NHL development [33, 34]. In addition, cytotoxic T cells, particularly activated CD8 T cells, are markedly elevated in diffuse large B-cell lymphoma, a major subtype among HIV-associated NHL [28]. Furthermore, HHV-8 during multicentric Castelman disease and EBV induce high virus-specific CD8 T-cell responses [35]. Such strong stimulation before cell proliferation may explain the positive association between high CD8 counts and NHL risk.

PLWH With CD4 ≥ 500/mm³

Early cART recommendations have resulted in changes in the characteristics of the HIV population, with a growing proportion of PLWH with CD4 ≥ 500/mm³ [22, 24]. CD4/CD8 ratio restoration remains difficult in this population because of persisting high CD8 counts [36–38]. It has been reported that 25% of PLWH starting cART with CD4 ≥ 500/mm³ still had CD8 > 1000/mm³ after 8 years of suppressive cART [38]. As previously shown, we confirmed that a large proportion (37%) of PLWH with CD4 ≥ 500/mm³ at baseline still had a CD4/CD8 < 1 after 5 years of suppressed viremia [39].

The association of a low CD4/CD8 ratio with morbidity has already been reported in subjects with CD4 ≥ 500/mm³, and maintaining a low CD4/CD8 ratio was associated with CD8 T-cell activation [10]. Moreover, a CD8 count above 1000/mm³ may be a predictor of new AIDS-defining events among PLWH with CD4 > 500/mm³ [40]. In our study, the CD4/CD8 ratio and CD8 count appeared to be particularly relevant markers in the context of high CD4 ≥ 500/mm³ to identify patients with increased risk of developing KS and NHL.

CONCLUSION

HIV-related morbidities still occur in PLWH despite suppressed viremia and relative immune restoration. Thus, biomarkers other than the traditional predictive CD4 count may be useful to identify the subset of individuals with a higher risk of morbidities. Previous studies have shown that the CD4/CD8 ratio may help to better monitor immune restoration and identify PLWH with a higher risk of non-AIDS-defining events, despite efficient cART. Here we showed that PLWH with low CD4/CD8 ratios or high CD8 counts, despite efficient cART, had a higher risk of developing KS and NHL, respectively, and that these markers, routinely assessed and easily available, may be particularly relevant in PLWH with CD4 ≥ 500/mm³. Closer clinical monitoring could be beneficial for PLWH with persisting low ratios despite efficient cART. However, further studies are needed to clarify the impact of the CD4/CD8 ratio on the risk of other HIV-related diseases, to know which specific preventives measures could be implemented.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We would like to thank Fabrice Carrat and Clovis Lusivika Zinga (Sorbonne Université, INSERM, Institut Pierre Louis d’Epidémiologie et de Santé Publique, Paris, France; Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Saint-Antoine, Unité de Santé Publique, Paris, France) for their help with the analysis.

Financial support. The COHERE study group has received unrestricted funding from: Agence Nationale de Recherches sur le SIDA et les Hépatites Virales (ANRS), France; HIV Monitoring Foundation, The Netherlands; and the Augustinus Foundation, Denmark. The research leading to these results has received funding from the European Union Seventh Framework Programme (grant number FP7/2007–2013) under EuroCoord grant agreement number 260694. A list of the funders of the participating cohorts can be found at www.COHERE.org.

Potential conflicts of interest. F. C. reports personal fees from Gilead France, personal fees from Janssen France, outside the submitted work. L. W. reports personal fees from Merck Sharp Dhome (MSD), other from ViiV Healthcare, outside the submitted work. A. W. reports personal fees, honoraria, and speaker fees from Gilead Sciences, ViiV Healthcare, Janssen, and MSD; and research grants paid to institution from Janssen, ViiV Healthcare, Gilead Sciences, BMS, and MSD, outside the submitted work. J. M. M. reports grants (IDIBAPS during 2017–21) and personal fees from AbbVie, Angelini, Contrafect, Cubist, Genentech, Gilead Sciences, Jansen, Medtronic, MSD, Novartis, Pfizer, and ViiV Healthcare, outside the submitted work. F. B. reports grants from Janssen, personal fees from ViiV Healthcare, Gilead, MSD, Janssen, outside the submitted work. P. R. reports grants from Gilead Sciences, grants from ViiV Healthcare, grants from Merck & Co, other from Gilead Sciences, Scientific Advisory Board participation, and honorarium paid to institution from ViiV Healthcare, Merck & Co, and Teva Pharmaceutical Industries, outside the submitted work. N. T. reports personal fees and other from Gilead, other from ViiV, personal fees, and other from Merck, outside the submitted work. L. M. reports grants from ANRS during the conduct of the study. C. B. reports personal fees from AbbVie, Gilead, Janssen, MSD, ViiV, grants from NEAT ID, Dt. Leberstiftung, DZIF, Hector Stiftung, outside the submitted work. A. A. reports grants, personal fees, and nonfinancial support from Gilead Sciences Inc, grants and personal fees from Janssen Cilag, grants, personal fees, and nonfinancial support from ViiV Healthcare, grants, and personal fees from MSD, outside the submitted work. L. W. reports grants from ANRS, during the conduct of the study. D. C. reports personal fees from Merck Switzerland, grants and personal fees from Janssen France, personal fees from Gilead France, grants and personal fees from MSD France, outside the submitted work. All other authors report no potential conflicts.. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Working Group for the Collaboration of Observational HIV Epidemiological Research Europe (COHERE) in EuroCoord

Steering Committee−Contributing Cohorts: Ali Judd (AALPHI), Robert Zangerle (AHIVCOS), Giota Touloumi (AMACS), Josiane Warszawski (ANRS CO1 EPF/ANRS CO11 OBSERVATOIRE EPF), Laurence Meyer (ANRS CO2 SEROCO), François Dabis (ANRS CO3 AQUITAINE), Murielle Mary Krause (ANRS CO4 FHDH), Jade Ghosn (ANRS CO6 PRIMO), Catherine Leport (ANRS CO8 COPILOTE), Linda Wittkop (ANRS CO13 HEPAVIH), Peter Reiss (ATHENA), Ferdinand Wit (ATHENA), Maria Prins (CASCADE), Heiner Bucher (CASCADE), Diana Gibb (CHIPS), Gerd Fätkenheuer (Cologne-Bonn), Julia Del Amo (CoRIS), Niels Obel (Danish HIV Cohort), Claire Thorne (ECS, NSHPC), Amanda Mocroft (EuroSIDA), Ole Kirk (EuroSIDA), Christoph Stephan (Frankfurt), Santiago Pérez-Hoyos (GEMES-Haemo), Osamah Hamouda (German ClinSurv), Barbara Bartmeyer (German ClinSurv), Nikoloz Chkhartishvili (Georgian National HIV/AIDS), Antoni Noguera-Julian (CORISPE-cat), Andrea Antinori (ICC), Antonella d’Arminio Monforte (ICONA), Norbert Brockmeyer (KOMPNET), Luis Prieto (Madrid PMTCT Cohort), Pablo Rojo Conejo (CORISPES-Madrid), Antoni Soriano-Arandes (NENEXP), Manuel Battegay (SHCS), Roger Kouyos (SHCS), Cristina Mussini (Modena Cohort), Jordi Casabona (PISCIS), Jose M. Miró (PISCIS), Antonella Castagna (San Raffaele), Deborah Konopnick (St. Pierre Cohort), Tessa Goetghebuer (St Pierre Paediatric Cohort), Anders Sönnerborg (Swedish InfCare), Carlo Torti (The Italian Master Cohort), Caroline Sabin (UK CHIC), Ramon Teira (VACH), Myriam Garrido (VACH). David Haerry (European AIDS Treatment Group).

Executive Committee: Stéphane de Wit (Chair, St. Pierre University Hospital), Jose Mª Miró (PISCIS), Dominique Costagliola (FHDH), Antonella d’Arminio-Monforte (ICONA), Antonella Castagna (San Raffaele), Julia del Amo (CoRIS), Amanda Mocroft (EuroSida), Dorthe Raben (Head, Copenhagen Regional Coordinating Centre), Geneviève Chêne (Head, Bordeaux Regional Coordinating Centre). Pediatric Cohort Representatives: Ali Judd, Pablo Rojo Conejo.

Regional Coordinating Centers: Bordeaux RCC: Diana Barger, Christine Schwimmer, Monique Termote, Linda Wittkop; Copenhagen RCC: Casper M. Frederiksen, Dorthe Raben, Rikke Salbøl Brandt.

Project Leads and Statisticians: Juan Berenguer, Julia Bohlius, Vincent Bouteloup, Heiner Bucher, Alessandro Cozzi-Lepri, François Dabis, Antonella d’Arminio Monforte, Mary-Anne Davies, Julia del Amo, Maria Dorrucci, David Dunn, Matthias Egger, Hansjakob Furrer, Marguerite Guiguet, Sophie Grabar, Ali Judd, Ole Kirk, Olivier Lambotte, Valériane Leroy, Sara Lodi, Sophie Matheron, Laurence Meyer, Jose Mª Miró, Amanda Mocroft, Susana Monge, Fumiyo Nakagawa, Roger Paredes, Andrew Phillips, Massimo Puoti, Eliane Rohner, Michael Schomaker, Colette Smit, Jonathan Sterne, Rodolphe Thiebaut, Claire Thorne, Carlo Torti, Marc van der Valk, Linda Wittkop.