Herpes zoster as a missed opportunity for HIV diagnosis: a real-world analysis of testing practices

Abstract

Background

Herpes zoster (HZ) is a recognized HIV indicator condition, and international guidelines recommend routine HIV testing in such cases. We aimed to evaluate HIV testing rates, HIV positivity, missed diagnostic opportunities, and the HIV late diagnosis rate among adults diagnosed with HZ.

Methods

This retrospective cohort study included adult patients (≥ 18 years) diagnosed with HZ between January 2022 and December 2024, at a tertiary referral center. Patients with a prior diagnosis of HIV infection were excluded. The study recorded demographic and clinical characteristics, as well as HIV testing status. HIV positivity rates among patients who underwent HIV testing were evaluated, alongside the rate of subsequent HIV diagnoses in patients who were initially not tested.

Results

A total of 2,446 patients with HZ were included. HIV testing was requested in 19.5% of patients at the time of HZ diagnosis. Among those tested, HIV positivity was 1.68%. Among patients not tested at baseline, 0.41% were subsequently diagnosed with HIV. The ARD between tested and non-tested groups was 1.26%. IPW analysis estimated an overall HIV prevalence of 1.78%, corresponding to approximately 44 expected cases; however, only 8 were diagnosed at initial presentation, indicating that a substantial proportion of HIV infections may not have been identified at the first clinical encounter.

Conclusion

Herpes zoster represents a significant missed opportunity for early HIV diagnosis. The low HIV testing rates and selective testing practices may contribute to late HIV diagnosis and missed opportunities for timely linkage to care. These findings support routine indicator condition–guided HIV testing to improve early detection and outcomes.

Background

Herpes zoster (HZ) is a disease resulting from the reactivation of a latent varicella-zoster virus infection, typically characterized by painful vesicular rashes. Key determinants in the development of HZ include age, immune system status, and underlying chronic conditions [1]. HIV infection is one of the strongest immunological risk factors for the development of HZ [2, 3].

According to UNAIDS 2024 report, 630 000 people died from AIDS-related illnesses [4]. Late diagnosis of HIV infection leads to increased AIDS-related disease, worse clinical outcomes, higher healthcare burden, and ongoing community transmission due to delayed antiretroviral therapy (ART). Consequently, early HIV diagnosis, transmission control, and improved access to treatment are critical objectives of global public health strategies [5]. The HIDES I study has shown that the systematic application of HIV testing in HZ cases—classified as an HIV indicator disease—substantially increases the rate of new diagnoses [6]. International guidelines, particularly those issued by EuroTEST, recommend HIV testing for all indicator diseases, including HZ, in countries in which the community HIV prevalence exceeds 0.1% [7]. Despite international recommendations, the practice of HIV testing among patients diagnosed with HZ remains inconsistent in real-world clinical settings.

Although the HIV prevalence in Türkiye ranges from 0.1% to 0.3%, studies indicate that the proportion of undiagnosed HIV cases remains notably high [8–10]. Therefore, in countries such as Türkiye, where HIV prevalence is low but the rate of undiagnosed cases is high, HIV testing rates for indicator diseases become particularly crucial.

This study aims to evaluate the rate of HIV testing, the HIV positivity rate, and the frequency of subsequent HIV diagnoses in patients with HZ who were not initially tested for HIV.

Materials and methods

This retrospective study was conducted at the largest referral hospital in Istanbul, which serves a population of 16 million. Adult patients aged 18 and older diagnosed with HZ between January 1, 2022, and December 31, 2024, at our hospital were evaluated. Patients were identified via the hospital’s electronic medical records system using ICD-10 codes (B02.0–B02.9). Although pediatric patients are also followed at our institution, this study was intentionally designed to include only adult patients. The epidemiology, clinical presentation, HIV testing practices, risk factors, and healthcare pathways of HZ and HIV infection differ substantially between pediatric and adult populations. Therefore, to ensure a more homogeneous study population and improve the validity of the analyses, only patients aged ≥ 18 years were included.

In our study, whether or not an HIV test was requested during the follow-up of patients diagnosed with HZ at our hospital was retrospectively reviewed using the electronic medical records system. Those with previously known HIV-positive status were excluded from the study. For each patient, data on age, gender, admission date, referring clinic (dermatology, infectious diseases, neurology, etc.), form of HZ (localized or complicated), comorbidities (diabetes mellitus, malignancy, autoimmune diseases, asthma/chronic obstructive pulmonary disease, chronic kidney failure), use of immunosuppressive medications, and anti-HIV testing status were recorded. Ward type was categorized as specialty outpatient clinic, general inpatient ward, and emergency department. Patients who had not been tested for HIV at the time of HZ diagnosis but were later diagnosed with HIV during subsequent visits were also included in the follow-up. The CD4 + T cell counts of patients diagnosed with HIV were measured at the time of diagnosis using flow cytometry. Late diagnosis of HIV infection was defined as individuals with CD4 + T cell counts below 350 cells/µL [11]. Complicated HZ cases were defined as those with ophthalmic involvement, Ramsay Hunt syndrome, or neurological complications (meningitis, encephalitis). Localized HZ cases included those with limited dermatomal distribution and no complications.

Primary endpoint

• Rate of HIV testing among patients diagnosed with HZ.

Secondary endpoints

• HIV positivity rate among tested patients.

• Rate of subsequent HIV diagnoses in initially untested patients.

• Rate of late diagnosis.

• Factors associated with HIV testing.

After organizing all the data, it was transferred to IBM SPSS Statistics 26.0 for analysis. Descriptive statistics for categorical variables were presented as frequency and percentage, and for continuous variables, as mean ± standard deviation or median (minimum–maximum). For comparisons between groups, the chi-square or Fisher’s exact test was used, with a p-value of < 0.05 considered statistically significant. For multivariable analysis, ward type was dichotomized as emergency department versus other clinical settings because of the markedly lower testing rates observed in emergency care. The HIV positivity rate in tested patients and the rate of subsequent HIV diagnoses in non-tested patients were compared as independent ratios. The difference between these two ratios was calculated as the absolute risk difference (ARD). The 95% confidence interval for ARD was computed using the Newcombe method.

In addition to the ARD analysis, inverse probability weighting (IPW) was applied to adjust for potential selection bias in HIV testing and to estimate HIV prevalence under a hypothetical universal testing scenario. Patients were stratified according to clinical presentation (complicated vs. localized zoster), and the probability of HIV testing was calculated separately for each subgroup.

These weights were applied to HIV-positive cases identified among tested individuals to estimate the weighted number of cases under a scenario of uniform testing. The weighted estimates were then used to calculate the adjusted HIV prevalence in the overall cohort. However, residual confounding due to unmeasured factors cannot be fully excluded. Of note, the ARD and IPW-derived estimates should be interpreted as measures of diagnostic yield rather than causal effect estimates.

Results

A total of 2,446 patients with HZ were included in the analysis after excluding 15 patients with known HIV infection (Figure 1). The mean age was 52.6 ± 17.6 years, with 46% of the patients being male.In this study, anti-HIV testing was requested for 478 (19.5%) patients at the time of diagnosis of HZ. Table 1 presents the rate of anti-HIV testing and the associated results. The HIV testing rate was 52.7% in patients with complicated forms and 16.8% in those with localized forms.

Fig. 1.

Study flow diagram of Herpes Zoster patients and HIV testing

Table 1.

HIV testing rates and HIV positivity according to sex and clinical presentation among patients with herpes zoster

Variable	Total (n)	Test requested (n, %)	HIV positivity (n, %)
Overall Total	2.446	478 (19.5)	8 (1.68)
Male	1.126	227 (20.2)	7 (3.1)
Female	1.320	251 (19.0)	1 (0.4)
Complicated Form of HZ	186	98 (52.7)	1(1)
Localized Form of HZ	2260	380 (16.8)	7 (1.8)

HZ, herpes zoster

Of the 478 patients who underwent testing, HIV positivity was identified in 8 patients (1.68%; 95% CI 0.83–3.31) (7 males, 1 female; mean age 43.6 ± 9.2 years). Three of these patients (37.5%) had a late diagnosis. The median CD4 + T cell count of these patients was 371/mm³ (min–max: 110–1788).

Among patients who were not tested for HIV at the time of HZ diagnosis, eight individuals (0.41%; 95% CI 0.21–0.81) subsequently presented to our clinic and were diagnosed with HIV infection. The median duration of later diagnosis (the interval between HZ diagnosis and HIV diagnosis) was 7.5 months (min–max: 2–14). All eight patients who were later diagnosed with HIV were male, with a mean age of 43.0 ± 11.1 years. All of these patients had localized form HZ lesions, and none had accompanying risk factors. Of these patients, six (75%) met the criteria for late diagnosis, and their median CD4 + T cell count was 203/mm³ (min–max: 50–938).

In terms of clinical characteristics, 92.4% of the cases presented with localized HZ, while 7.6% (n = 186) had the complicated form. Among the 186 complicated cases, ophthalmic involvement was noted in 131, Ramsay Hunt syndrome (ear involvement) in 13, and neurological complications (meningitis or encephalitis) in 57 Cases with multiple involvements were rare (eye + ear in three, eye + neurological in 11, ear + neurological in one), and no cases of triple involvement were observed. The incidence of complicated forms did not exhibit a statistically significant association with gender (p = 0.205). The most common comorbidity was diabetes mellitus (25.2%), and more than half of the patients were aged 50 years or older.

In the bivariate analysis between HIV testing request status and clinical variables, statistically significant associations were observed with age ≥ 50, cancer history, use of immunosuppressive medications, and complicated HZ (p < 0.05). Table 2 illustrates the analysis of the relationship between demographic and clinical characteristics and the request for anti-HIV testing.

Table 2.

Clinical and demographic factors associated with HIV test requests in patients presenting with herpes zoster

Variable	Condition	Test requested (n)	Test not requested(n)	Total	p-value
Age (years)	< 50	163	836	999	0.001*
	≥ 50	315	1132	1447
Gender	Female	251	1069	1320	0.509
	Male	227	899	1126
Cancer		88	147	235	< 0.001*
Use of immunosuppressive drugs		42	93	135	< 0.001*
Complicated (Severe HZ)		98	88	186	< 0.001*
Ward Type	Speciality Outpatient	226	1320	1546	< 0.001*
	General Ward (Inpatient)	204	37	241
	Emergency Room	48	611	659

HZ, herpes zoster. Comparisons performed using chi-square or Fisher’s exact test. * p < 0.05

In the multivariate logistic regression analysis (Table 3), ward type was a strong independent predictor of HIV testing. Compared to speciality outpatient settings, patients admitted to general inpatient wards were significantly more likely to be tested (OR = 25.93; 95% CI: 17.47–38.48; p < 0.001), whereas emergency room attendances were associated with substantially lower testing odds (OR = 0.47; 95% CI: 0.34–0.65; p < 0.001). Among clinical variables, cancer (OR = 1.90; 95% CI: 1.33–2.73; p < 0.001) and complicated HZ (OR = 1.57; 95% CI: 1.02–2.43; p = 0.043) remained independent predictors after adjustment for ward type. Immunosuppressive drug use (OR = 1.30; 95% CI: 0.73–2.29; p = 0.373) and age ≥ 50 (OR = 0.85; 95% CI: 0.67–1.10; p = 0.218) were not retained as independent factors in the multivariate model.

Table 3.

Factors independently associated with HIV testing among patients with herpes zoster: multivariable logistic regression analysis

Variable	OR	%95 CI Lower	%95 CI Upper	p-value
Age ≥ 50	0.85	0.67	1.10	0.218
Cancer	1.90	1.33	2.73	< 0.001*
Use of immunosuppressive drugs	1.30	0.73	2.29	0.373
Complicated (Severe) HZ	1.57	1.02	2.43	0.043*
Emergency Room vs. others	0.47	0.34	0.65	< 0.001*

HZ, herpes zoster. Ward type was dichotomized as emergency department versus other clinical settings. Odds ratios > 1 indicate increased likelihood of HIV test request. OR: Odds Ratio; CI: Confidence Interval. * p < 0.05

The ARD between the HIV positivity rate in tested patients and the rate of late diagnosis in non-tested patients was 1.26%, with a 95% confidence interval (Newcombe) of 0.08–2.44. To mitigate potential selection bias in HIV testing, IPW analysis was conducted. The weighted HIV prevalence in the overall cohort was estimated to be 1.78%, which corresponds to approximately 44 expected HIV infections in a hypothetical universal testing scenario. However, only 8 HIV infections were diagnosed at the time of HZ presentation, indicating that roughly 36 HIV infections may have been missed during the initial clinical assessment. Although 8 of these patients were subsequently diagnosed during follow-up, an estimated 28 HIV infections may have remained undiagnosed throughout the study period. This, approximately 64% of expected HIV-positive cases may not have been identified at the initial clinical encounter.

The HIV testing request rate varied significantly across clinics. The highest testing rates were observed in dermatology, ophthalmology, and internal medicine departments, while the lowest rate was recorded in the emergency department. The HIV testing request rates across the clinics are presented in Table 4.

Table 4.

HIV testing request rates across clinics

Clinics	Total HZ patients (n)	Test requested (n)	Test request rate (%)
Dermatology	1 559	383	24,6
Ophthalmology	45	11	24,4
Internal Medicine	41	10	24,4
Infectious Diseases	46	10	21,7
General Practitioner	16	3	18,8
Neurology	29	3	10,3
Emergency Department	659	48	7,3
Other Specialties	51	10	19,6

Discussion

This study aimed to assess HIV testing rates, test outcomes, and subsequent HIV diagnoses among patients diagnosed with HZ. Despite the well-established association between HZ and HIV, HIV testing remains suboptimal in routine clinical practice, leading to missed diagnostic opportunities and a higher rate of delayed HIV diagnoses, particularly in individuals without classical immunosuppressive risk factors.

In a meta-analysis by Marra et al., it was shown that HIV infection increases the risk of developing HZ by approximately threefold (RR = 3.22; 95% CI: 2.40–4.33) [1]. Ku et al. reported that the incidence of HZ in people living with HIV (PLWH) is 3–20 times higher compared to the general population [3]. Similarly, in a systematic review by McKay et al., the incidence of HZ in HIV-positive individuals ranged from 9.3 to 141 per 1,000 person-years, which is 3 to 20 times higher than in the general population [12]. These findings underscore the notion that HIV infection is a strong epidemiological and clinical predictor for the development of HZ.

The prevalence of HIV in HZ patients (2.3–4.9%) has been studied in a limited number of studies, with the reported number of cases generally being low [6, 13–15]. In our study, HIV prevalence in HZ patients tested at the time of diagnosis was found to be 1.68%. This rate is approximately 6 to 15 times higher than that of the general population in Türkiye (0.1–0.3%) [9].

In a population-based study conducted by Agustí et al. in Spain, HIV testing was requested in only 22.7% of cases with indicator diseases, and this rate decreased to 2.9% in the HZ subgroup [16]. In our study, the HIV testing request rate was 19.5%, and the HIV positivity rate was 1.68%. Moreover, 0.4% of patients who were not tested later received an HIV diagnosis. This discrepancy suggests that, although testing rates in Türkiye are relatively higher than those reported in Spain, the proportion of missed diagnoses remains considerable. While no HIV diagnoses were made in non-tested cases in Agustí et al.’s study, our study uniquely contributes by revealing the potential diagnostic gap (missed opportunity) in non-tested patients [16].

Among patients with HZ who underwent HIV testing, higher testing rates were observed in those with complicated forms. This suggests that clinicians may primarily associate HIV infection with complicated HZ or immunosuppressive conditions. By applying an inverse probability weighting approach, we demonstrated that the estimated burden of HIV infection in this population may be considerably higher than observed. The adjusted estimate suggests that a substantial proportion of HIV-positive cases may remain undiagnosed at the time of HZ presentation. Importantly, all patients subsequently diagnosed with HIV initially presented with localized HZ, which challenges the common assumption that only severe or disseminated forms of the disease warrant HIV testing. This finding emphasizes the importance of considering HIV testing even in clinically mild presentations.

In a study by Yaylalı et al., modeling the dynamics of the HIV epidemic in Türkiye, it was projected that by 2020, 63% of HIV infections remained undiagnosed, and if current diagnosis rates are maintained, the number of PLWH could reach 2.4 million by 2040 [8, 10]. The model demonstrated that increasing the diagnosis rate to 90% would reduce new infections by 97%. However, the low HIV testing request rates observed in our study strongly support the persistence of the national testing gap in Türkiye, highlighting the necessity for the expansion of indicator disease-based testing strategies at a national level. These findings are particularly relevant for middle-income countries such as Türkiye, where healthcare utilization is high but structured HIV screening strategies remain limited.

In our study, 8 of the 1,968 patients who were not tested at the time of HZ diagnosis were subsequently diagnosed with HIV. Had testing been performed at presentation, these eight patients could have been diagnosed an average of 7.5 months earlier. Late HIV diagnosis is associated with increased morbidity and mortality [17]. In our cohort, 3 of the 8 patients tested at the time of HZ diagnosis (37.5%) presented late, whereas 6 of the 8 patients who were not initially tested (75%) experienced late diagnosis. These findings suggest that delayed HIV diagnosis may adversely affect clinical outcomes.

The low HIV testing rate observed in our study may reflect physician-related barriers. Clinicians may underestimate HZ as an HIV indicator condition, particularly in older patients or those with accompanying comorbidities. Although HIV testing is widely accessible and covered by the national healthcare system in Türkiye, time constraints during clinical encounters, and persistent HIV-related stigma may adversely affect the routine implementation of HIV testing.

From a public health perspective, late diagnosis represents a critical deficiency in the first pillar of the United Nations Joint Programme on HIV/AIDS 95-95-95 targets [18]. Undiagnosed HIV infections not only result in delays in treatment at the individual level but also contribute to the continued transmission of the virus within the community.

The IPW analysis conducted in this study indicates that HIV testing in HZ patients was not systematically performed, but rather selectively, which may have led to an underestimation of the estimated HIV burden. This finding underscores the persistence of undiagnosed cases within the system. While IPW reduced bias related to measured clinical factors, it could not account for unmeasured determinants of HIV risk and testing behavior.It further highlights the importance of indicator disease-based testing strategies in HIV screening efforts.

Our study identified significant differences in HIV testing request rates across various clinical settings, suggesting a marked heterogeneity in testing behavior. This variability likely reflects differences in HIV awareness among clinicians across specialties. García-García et al. demonstrated that a brief educational intervention significantly improved clinicians’ HIV screening practices, leading to a notable increase in both the number of tests requested and the number of new HIV diagnoses [19].

In our study, higher testing rates were observed in the infectious diseases and dermatology departments, while the lowest rates were recorded in the emergency care settings. Emergency departments in Türkiye are under significant pressure due to high patient volumes; a study conducted in the country reports that annual emergency department visits can reach up to twice the total population (approximately 1,300–1,550 visits per 1,000 people) [20]. This high volume of visits shifts the priority of emergency physicians toward managing life-threatening conditions, leading to missed diagnostic opportunities, such as HIV testing based on indicator diseases, which require heightened clinical awareness. The fact that four out of the eight missed HIV diagnoses in our study originated from the emergency department is a direct reflection of this systemic burden. Therefore, the implementation of indicator disease-based testing protocols in emergency settings should be regarded as a priority national public health intervention to detect undiagnosed HIV cases early and break the transmission chain. This underscores the emergency department as a critical yet underutilized setting for HIV case identification.

Meta-analyses in the existing literature clearly demonstrate that HIV infection significantly increases the risk of HZ development [12, 21]. However, our study shows that despite this elevated risk, this relationship is often overlooked in clinical practice, with HIV testing rates remaining low and, consequently, significant delays in diagnosis.

The limitations of this study include its retrospective design and single-center setting, which may limit the generalizability of the findings. HIV testing was performed at the discretion of treating physicians rather than through a standardized protocol, introducing the possibility of indication and selection bias. Although IPW was applied to reduce bias related to non-random HIV testing practices, residual confounding due to unmeasured clinical, behavioral, and socioeconomic factors cannot be completely excluded. In addition, HIV testing status was determined based on laboratory records available within the institutional hospital information system. Because HIV testing was not systematically performed in all patients, the estimated proportion of potentially missed HIV diagnoses should be interpreted cautiously.

Future studies should incorporate prospective, multicenter cohorts to better understand testing behaviors and missed diagnostic opportunities. Intervention-based studies evaluating structured HIV testing strategies may further enhance early diagnosis and support progress toward the 95-95-95 targets.

In conclusion, HZ represents a missed opportunity for timely HIV diagnosis in routine clinical practice. Routine HIV testing in adults presenting with HZ, regardless of clinical severity, may be considered an important component of standard care to improve timely diagnosis, facilitate prompt linkage to care and treatment, better characterize the HIV epidemic in Türkiye, and strengthen comprehensive HIV prevention efforts.

Abbreviations

HZ

Herpes zoster

ARD

Absolute risk difference

IPW

Inverse probability weighting

CI

Confidence ınterval

PLWH

People living with HIV

OR

Odds ratio

Author contributions

Design of study: RK, EÇ, ÖAA, MMK. Acquisition, analysis, and interpretation of data: RK, EÇ, AG, MOG,ÖAA, MMK. Drafting and reviewing the work: RK, EÇ, AG, MOG, ÖAA, MMK Approval of the final version: RK, EÇ, AG, MOG, MMK,ÖAA.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Institutional Ethics Committee of Başakşehir Çam and Sakura City Hospital (approval number: KAEK/271/2025). The study was conducted in accordance with the Declaration of Helsinki. As this was a retrospective cohort study based on routinely collected clinical data, informed consent was waived by the ethics committee.

Consent for publication

Not applicable, as the study analysed retrospectively collected and anonymized clinical data.

Competing interests

The authors declare no competing interests.