Late Diagnosis of HIV Infection: the Role of Age and Gender

Abstract

Background

Late diagnosis of HIV infection is detrimental to infected persons and to the public health. The objective of this study was to identify factors associated with late diagnosis of HIV infection, defined as an initial CD4 T lymphocyte count < 200 cells/µL, in a cohort of recently diagnosed persons. Additionally, we evaluated factors associated with HIV infection being diagnosed during hospitalization.

Design

Cross-sectional study of a university-based HIV clinic in the southeastern U.S. Patients with newly diagnosed HIV infection evaluated at the Duke University HIV clinic between October 2002 and August 2004 were included in this analysis. Socio-demographic variables, site of HIV diagnosis, opportunistic infections present at diagnosis, initial CD4 T lymphocyte count and initial HIV RNA level were recorded for study subjects.

Results

49% of subjects met the immunologic definition of AIDS at the time of HIV diagnosis (CD4 count < 200 cells/µL). In multivariable logistic regression analyses, older patients were more likely to be diagnosed with a CD4 count < 200 cells/µL (OR=1.72, 95% CI= 1.12,2.64, p=0.01), and older patients (OR=1.79, 95% CI= 1.07,3.12, p=0.03) and women (OR=6.74, 95% CI= 2.08,21.81, p=0.001) were more likely to be diagnosed during hospitalization.

Conclusions

Late diagnosis of HIV infection is a considerable problem, particularly for older patients. Inpatient diagnosis of HIV infection is significantly more common among women and older patients. Improved HIV testing strategies may allow for more timely diagnosis of HIV infection, which may benefit both the infected individual and society.

Introduction

Highly active antiretroviral therapy (HAART) has proven effective in reducing morbidity and mortality associated with human immunodeficiency virus (HIV) infection.¹ Despite increased availability of HAART in the United States many HIV-infected persons do not fully benefit from therapy due to late diagnosis. Approximately 25% of the estimated 1.1 million persons living with HIV in the United States are unaware they are HIV-infected.² As a result, these persons do not achieve full benefit from life-prolonging antiretroviral and prophylactic medications. Recent data from the Centers for Disease Control and Prevention (CDC) indicate that approximately half the persons meeting the case definition for AIDS had been diagnosed with HIV infection less than a year before their AIDS diagnosis.³ Of all persons diagnosed with HIV infection in the U.S. in 2001, 39% progressed to AIDS within 12 months of their positive HIV test.⁴

This late diagnosis of HIV infection is detrimental both to the individual and to society. HIV-infected patients whose CD4 counts are below 200 cells/µL experience increased morbidity and mortality.⁵ Additionally, diminished responses to antiretroviral therapy and higher health care expenditures are the norm for HIV-infected persons with lower CD4 counts.^5–7 There has been a recent focus on incorporating prevention of transmission strategies into the medical care of persons living with HIV, and trials have shown that focusing prevention counseling on HIV-infected persons reduces rates of high-risk sexual behaviors that might transmit HIV.^8,9 Thus, early diagnosis of HIV infection has become increasingly important on many levels.

Studies assessing factors associated with late presentation for HIV care suggest older age and male gender are associated with late entry to care in some studies but not in others.^10–12 Most studies have focused on patients living in large urban settings in the Northeast and West. However, the South has the highest rates of new HIV infections and is home to 39% of persons living with AIDS in the U.S.¹³ Additionally, the socio-demographic characteristics of HIV-infected patients in the South differ from those observed in other regions of the country; over two thirds of patients diagnosed with AIDS in 2003 and residing in non-metropolitan areas live in the South.¹⁴ A study done over a decade ago showed that only 12% of recently diagnosed HIV-infected patients in South Carolina had CD4 T lymphocyte counts < 200 cells/µL,¹⁵ suggesting late diagnosis was relatively infrequent. Much has changed in the past 10 years regarding features of persons living with HIV/AIDS, including increased prevalence rates among African-Americans, Latinos, and women.¹³ Further, public health funding has declined over this time. Given the importance of the timing of HIV diagnosis for the individual and the public health, an assessment of HIV diagnosis in the current era seemed appropriate.

To determine characteristics of persons with recently diagnosed HIV infection in one area of the South we reviewed medical records to correlate the timing and location of HIV diagnosis with various socio-demographic factors.

Methods

Study design

We evaluated the medical records of all clinic patients initially presenting for care at the Duke University HIV clinic between October 2002 and August 2004. Only persons whose HIV infection had been diagnosed within 6 months of presentation to the clinic were included in this analysis. Abstracted data included age, gender, race/ethnicity, reported mode of HIV acquisition, insurance status, 5-digit zip code, site of diagnosis (inpatient vs. outpatient), presence of an opportunistic infection at diagnosis, initial CD4 T lymphocyte count, and initial HIV RNA level. Non-Caucasians were classified as racial/ethnic minorities. Consistent with the CDC and Office of Management and Budget classification, persons not residing in a metropolitan statistical area (population less than 50,000) were considered to reside in non-metropolitan locales.^13,14,16 Age was categorized in 10-year increments. The study was approved by the Duke University Institutional Review Board.

Statistical analysis

Univariable analyses were performed on all study variables to ensure that distributional assumptions were met. Bivariable analyses were done using Pearson’s chi-square test for dichotomous variables and logistic regression for continuous variables. The study had two primary outcome measures; a CD4 count < 200 cells/µL (diagnostic of immunologic AIDS) at the time of presentation, and an initial diagnosis of HIV infection while an inpatient. Multivariable logistic regression analyses were performed for both outcome variables. Predictor variables were identified a priori and included age, gender, race/ethnicity, insurance status, and rural residence. All statistical analyses were performed using SAS E-Guide Version 2.05 for Windows (SAS Institute, Cary, NC). Two-sided p-values at the 0.05 level were used to determine statistical significance (α = 0.05).

Results

Data were available for 113 patients who met study inclusion criteria (Table 1). The study population was typical of the Duke University HIV clinic with small numbers of intravenous drug users (7%), and more than 25% women. Sixty-three percent of subjects were racial/ethnic minorities (87% African-American, 10% Latino), and nearly a third lived in non-metropolitan areas. Initial CD4 T lymphocyte counts were < 200 cells/µL in 49% of subjects and 35% were diagnosed during hospitalization.

Table 1.

Patient characteristics of 113 recently diagnosed HIV-infected subjects establishing care at the Duke University HIV Clinic.

Characteristic	Mean ± SD or Frequency (%)
Age (range 17–61 years)	36.1 ± 10
Female	31 (27.4%)
Race
Caucasian	35 (31.0%)
Minority	71 (62.8%)
Unknown	7 (6.2%)
Reported mode of HIV acquisition
Men who have sex with men	33 (29.2%)
Heterosexual contact	32 (28.3%)
Intravenous drug use	8 (7.1%)
Other/Unknown	40 (35.4%)
Uninsured	44 (38.9%)
Non-metropolitan	32 (28.3%)
Site of Diagnosis
Inpatient HIV diagnosis	39 (34.5%)
Outpatient HIV diagnosis	54 (47.8%)
Unknown	20 (17.7%)
Opportunistic Infection at diagnosis*	20 (17.9%)
HIV RNA > 100,000 copies/mL	66 (58.9%)
CD4 Lymphocyte count <200 cells/µL	55 (48.7%)

^*Opportunistic Infections: Pneumocystis carinii pneumonia (PCP) 9, Cryptococcal meningitis 5, Lymphoma 2, Cryptosporidial diarrhea 1, CNS toxoplasmosis 1, Progressive multifocal leukoencephalopathy 1, Kaposi’s sarcoma 1

Bivariable analyses

Older patients were significantly more likely to have an initial CD4 count < 200 cells/µL (p=0.005) (Table 2). Diagnosis of HIV infection during hospitalization was more common among older patients (p=0.02), females (p<0.001), and minorities (p=0.02) (Table 2).

Table 2.

Bivariable and multivariable logistic regression analyses. Variables associated with immunologic AIDS (CD4 count < 200 cells/µL) and inpatient HIV diagnosis among 113 recently diagnosed HIV-infected subjects establishing care at the Duke University HIV Clinic.

	CD4 count < 200 cells/µL (n=113)				Inpatient HIV Diagnosis (n=93)
Characteristic	Bivariable analysis		Multivariable analysis		Bivariable analysis		Multivariable analysis
	OR (95%CI)	p-value	OR (95%CI)	p-value	OR (95%CI)	p-value	OR (95%CI)	p-value
Age* (range 17–61 years)	1.74 (1.16,2.60)	0.005	1.72 (1.12,2.64)	0.01	1.63 (1.07,2.56)	0.02	1.79 (1.07,3.12)	0.03
Female	1.68 (0.73,3.88)	0.22	1.30 (0.51,3.35)	0.59	6.71 (2.52,17.87)	<0.001	6.74 (2.08,21.81)	0.001
Minority	1.37 (0.61,3.08)	0.45	1.56 (0.63,3.87)	0.34	3.27 (1.20,8.87)	0.02	2.63 (0.82,8.43)	0.10
Uninsured	0.94 (0.44,2.00)	0.87	1.06 (0.46,2.45)	0.89	1.71 (0.74,4.00)	0.21	2.99 (0.99,9.00)	0.05
Non-metropolitan	1.28 (0.57,2.91)	0.55	1.19 (0.49,2.87)	0.71	1.58 (0.63,3.93)	0.33	1.32 (0.44,3.99)	0.61

^*Odds ratios for 10 year increase in age

Multivariable analyses

Older patients (AOR=1.72, 95% CI= 1.12,2.64, p=0.01) were more likely to be diagnosed with a CD4 count < 200 cells/µL (Table 2). Diagnosis of HIV infection during hospitalization was more common among older patients (AOR=1.79, 95% CI= 1.07,3.12, p=0.03) and females (AOR=6.74, 95% CI= 2.08,21.81, p=0.001) while lack of insurance was of borderline statistical significance (AOR=2.99, 95% CI=0.99,9.00, p=0.05) (Table 2).

Discussion

Despite efforts to increase HIV awareness, half the recently diagnosed HIV-infected patients evaluated at a university-based HIV clinic in North Carolina had initial CD4 T lymphocyte counts < 200 cells/µL compared to only 12% of those reported a decade earlier from South Carolina.¹⁵ Despite prevention and awareness campaigns, the problem of late diagnosis appears to have dramatically worsened over time. This may relate to changes in socio-demographic characteristics among persons living with HIV/AIDS, such that persons now at greatest risk are less likely to be tested due to issues of stigma and lower perceived risk. The ready availability of effective antiretroviral medications may also play a role by providing a sense of security to at risk populations thereby reducing the urgency with which they seek HIV testing. Further, declines in public health spending over this time may also contribute to this finding.

In addition to detrimental effects for the individual,^5–7 late diagnosis of HIV infection increases risk of transmission to others. Prevention counseling that can reduce risk behaviors and secondary infections is generally focused on persons in whom HIV infection is already known.^8,9 Therefore, patients diagnosed late have “missed out” on prevention programs for months to years while unaware of their infection. Moreover, antiretroviral therapy that reduces HIV RNA levels almost certainly reduces risk of transmission to sexual partners.¹⁷ The delay in initiating antiretroviral therapy due to late diagnosis thus likely generates more secondary HIV infections.

Testing based on perceived risk of HIV infection, rather than broader screening approaches, likely contributes to the problem. This study shows that older patients are more likely to be diagnosed both during hospitalization and with a CD4 count < 200 cells/µL. Additionally, women had a 6-fold increased likelihood of being diagnosed in the hospital. This suggests that older patients and women may have lower perceived risk for HIV infection by both themselves and by their providers, and are therefore less likely to be tested for HIV as outpatients. In multivariable logistic regression analysis, uninsured patients were also more likely to be diagnosed as inpatients, perhaps because uninsured patients have less access to non-emergent health care. Living in a non-metropolitan area, a characteristic of the HIV epidemic unique to the South, was surprisingly not associated with late diagnosis of HIV infection despite the reduced availability of medical services, and increased HIV stigma in rural communities.

As a single center study our findings may not be generalizable to other medical settings or regions of the country. It is also possible that advanced HIV-infected patients may be over-represented in this sample since sicker patients may be more likely to be referred to an academic medical center for care. However, our findings are consistent with other recently reported estimates, including a cohort of patients in a large health maintenance organization.^10,11 Given the relatively modest sample size it is possible that there was insufficient power to identify other important associations such as whether minorities and uninsured patients may be more likely diagnosed during hospitalization. Finally, as an observational study these data identify associations, but can not attribute causality.

These findings support the notion that late diagnosis of HIV infection remains a significant problem, particularly among older patients. Additionally, these data indicate diagnosis of HIV infection during hospitalization is disturbingly common, especially among older patients and women. Recent initiatives call for HIV testing to become a more routine part of medical care,^18,19 in part because of the problem of late HIV diagnosis substantiated by this study. Routine rather than risk-based HIV testing is advocated since high-risk behaviors are frequently not identified in primary care encounters.¹¹ Prospective studies assessing the impact of routine screening in various medical settings are ongoing and will be important in developing an evidence-based HIV screening policy.

Clinical significance.

• 49% of recently diagnosed HIV-infected patients had clinical and/or immunologic AIDS (CD4 T lymphocyte count < 200 cells/µL).

• Older patients were more likely to have an AIDS diagnosis at the time HIV was diagnosed.

• Older patients and women were more likely to be diagnosed with HIV during hospitalization.

• These findings support recent calls for more routine HIV testing in primary care settings.