Survival after an HIV diagnosis has improved considerably in industrialized countries since the introduction of highly active antiretroviral therapy (HAART) in 1996.1,2 As drug therapies improve and evolve,3 it is important to continue to provide the most up-to-date population-level estimates of survival and life expectancy among the HIV-infected population to understand the collective effectiveness of treatment therapies, services, and prevention programs that extend the lives of the HIV-infected population. Survival and life expectancy are important measures of disease burden that assist policy makers in allocating resources. Healthy People 2010 Objectives 13-15 and 13-16 address extending the years of life of persons diagnosed with HIV infection.4 The results provided here monitor the achievement of these objectives.
Quantitative measures of the improvement in life expectancy and average years of life lost (AYLL) after diagnosis using population-based surveillance data have been computed for other diseases but rarely used for HIV diagnosis in the United States. The AYLL (also called potential years of life lost) estimates the average time a person would be expected to live had she or he not died prematurely.5 It is also a measure of burden to the individual patient. In addition to being helpful to policy makers, knowledge of both of these measures could improve clinician communications with patients.6 The measures estimate how many remaining years of life a newly HIV-diagnosed person has to live and by how many years an HIV-diagnosed person's life would be shortened, on average, compared with a similar person in the general population.
A few studies in the post-HAART era have compared HIV-infected patients with the general population on life expectancy or mortality.7-15 All but two13,14 were cohort studies with patients receiving care and were primarily from European countries. All found continued excess mortality in HIV-infected populations compared with the general population. We used national population-based HIV surveillance data from 25 states to estimate the improvement in life expectancy and AYLL over a 10-year period compared with the general US population. Because diagnosis and mortality rates vary by race/ethnicity, we provide analyses by race/ethnicity.
We used US national HIV surveillance data from 25 states with name-based HIV surveillance since 1996. States included in the analysis were Alabama, Arizona, Arkansas, Colorado, Idaho, Indiana, Louisiana, Michigan, Minnesota, Mississippi, Missouri, New Jersey, Nevada, North Carolina, North Dakota, Ohio, Oklahoma, South Carolina, South Dakota, Tennessee, Utah, Virginia, West Virginia, Wisconsin, and Wyoming. HIV diagnoses among adults and adolescents (≥13 years) from 1996 through 2005 with follow-up through 2007 and reported to the Centers for Disease Control and Prevention (CDC) without identifying information through June 2008 were included. Deaths were included irrespective of cause of death. HIV surveillance data were collected as part of routine practices as mandated by state and local laws or regulations. The CDC determined that HIV surveillance is not a research activity and therefore does not require review by an institutional review board. The CDC used its Guidelines for Defining Public Health Research and Public Health Non-Research16 and Title 45 Part 46 of the Code of Federal Regulations17 to make its determination.
We modeled life expectancy after an HIV diagnosis using the life table approach.18 The period life table method estimates life expectancy of a hypothetical cohort by assuming that those at a given age in future years would have the same age-specific death rates as those at the same age in the current year. We used an exponential model to estimate the age-specific death rates. Because the death rate within 1 year after HIV diagnosis is quite different from that more than 1 year after HIV diagnosis, we estimated the age-specific death rates separately for death within 1 year after HIV diagnosis and for that more than 1 year after HIV diagnosis (Song R, PhD, Qin G, PhD, McDavid Harrison K, PhD, MPH, et al. Modeling survival after AIDS diagnosis based on National AIDS Surveillance Data, unpublished data, May 2009). In addition, data were adjusted for case and death reporting delay19 and missing risk factors.20
We then compared life expectancy of a person after HIV diagnosis with that of a person in the general population with the same age, sex, and race/ethnicity in the same calendar year using mortality data from the general population18 to estimate the AYLL after HIV diagnosis. US National Center for Health Statistics (NCHS) vital statistics data were used to prepare race/ethnic life tables using age-, sex-, race-, and calendar year-specific death rates, with race/ethnicity categories consistent with those in HIV surveillance data [white, black/African American (black), Hispanic/Latino (Hispanic)]. Numbers of American Indian/Alaska Native, Asian, and Native Hawaiian/Other Pacific Islander were small and produced unstable estimates; therefore, these groups' results were not presented. Numerators for death rates were obtained from mortality data from the NCHS for 2005.21 Population denominators for death rates were based on bridged-race intercensal estimates for 2005.22 The death rate was modeled as a function of age and estimated using cross-sectional data observed in a specific calendar year from 1996 to 2005.
The AYLL was calculated as the difference between life expectancy by age in the general and the HIV-diagnosed populations. Estimates of life expectancy by sex, race/ethnicity, CD4+ T-lymphocyte count at or within 6 months after diagnosis [AIDS opportunistic infection or CD4 count <200 cells/μL or percent <14 (stage 3), 200≤ CD4 count <500 cells/μL or 14≤ CD4 percent <27 (stage 2), CD4 count ≥500 or CD4 percent ≥27 (stage 1), and other or missing CD4 count],23 and HIV transmission category [male-to-male sex, injection drug use (IDU), male-to-male sex and IDU, and high-risk heterosexual contact (HC) (with a sex partner known to have, or to be at high risk for, HIV infection, eg, a man who has sex with men or an injection drug user)] are presented along with figures of AYLL by sex and race/ethnicity.
A total of 220,646 people were diagnosed with HIV in the 25 states from 1996 to 2005, and 10,366 deaths occurred among them by the end of 2007. Seventy-four percent of the infections occurred in males, 55% among blacks, 36% among whites, and 9% among Hispanics/Latinos (Table 1). Thirty-three percent had a first CD4 count <200 at or within 6 months after HIV diagnosis, 16% had a first CD4 count of at least 200 but not more than 499, 9% had a first CD4 count ≥500, and 42% had a missing CD4 count or 1 first measured more than 6 months after diagnosis. Among males, 67% were exposed through male-to-male sexual contact, 16% through IDU, 10% through HC, and 7% through male-to-male sexual contact and IDU. Seventy percent of females were exposed through HC and 29% through IDU; transmission category was unknown for 2%.
Average life expectancy after HIV diagnosis in 25 states increased from 10.5 years [95% confidence interval (CI): 10.4 to 10.6] to 22.5 years (95% CI: 22.2 to 22.7) from 1996 to 2005 (Table 2) and decreased with each increasing year of age at HIV diagnosis in 2005 (data not shown). Improvements in life expectancy were incremental each year except from 2000 to 2001, largely due to a decrease in life expectancy that year among females diagnosed with HIV infection. The largest 1-year increase was from 1996 to 1997 when HAART use was expanded. For each diagnosis year from 1997 to 2005, life expectancy was longer for females than for males and improved for both groups from 1996 to 2005 [12.6 (95% CI: 12.4 to 12.7) to 23.6 (95% CI: 23.0 to 24.2) years for females and 9.9 (95% CI: 9.8 to 10.0) to 22.0 (95% CI: 21.8 to 22.2) years for males] (Table 2).
Among males diagnosed in 2005, life expectancy was longest for white males, followed by Hispanic and then black males. Life expectancy improved less for black males [9.5 (95% CI: 9.4 to 9.7) to 19.9 (95% CI: 19.6 to 20.2) years] than for white males [10.3 (95% CI: 10.1 to 10.4) to 25.5 (95% CI: 24.9 to 26.1) years] and Hispanic males [10.1 (95% CI: 9.8 to 10.4) to 22.6 (95% CI: 21.9 to 23.3) years] from 1996 to 2005 (Table 3). Annual increases in life expectancy were experienced by black males, whereas among Hispanic males, improvement in life expectancy was not monotonic.
Due to a limited number of HIV diagnoses among white and Hispanic females, results of their estimated life expectancies are not stable. Overall, Hispanics and whites had similar life expectancy in the more recent years, and black females had the shortest life expectancy after HIV diagnosis. Improvements were not very significant for Hispanic and white females diagnosed in the last 5 years from 2001 to 2005. During this period, the average life expectancy was 20.6 years for black females (8-year increase from 1996), 22.1 years for white females (9-year increase), and 24.2 years for Hispanic females (13-year increase) (Table 3).
For each year from 1997 to 2004, males with male-to-male sex had the longest life expectancy among transmission categories. For those diagnosed in 2005, there was no significant difference among males in life expectancy between transmission categories male-to-male sex [28.3 (95% CI: 27.4 to 29.2) years] and male-to-male sex and IDU [24.6 (95% CI: 20.6 to 28.7) years] (Table 4). For females diagnosed in 2005 and in the HC transmission category had almost double the life expectancy [29.6 (95% CI: 27.9 to 31.3) years] of females in the IDU category [15.9 (95% CI: 15.1 to 16.6) years]. Each year after 1996, females in the HC category had roughly a 10-year longer life expectancy than those in the IDU category. Injection drug users, both male and female, had the worst life expectancy among all transmission categories for each year of the 10-year period.
Life expectancy increased almost 4-fold (5.5-19.4 years) among HIV-diagnosed persons from 1996 to 2005 with a first CD4 count at or within 6 months after diagnosis <200 (Table 2). Improvement in life expectancy over the 10-year period was even greater for persons with a CD4 count of <500 but at least 200 from 6.6 to 21.1 years.
Compared with the general US population (assuming that the standard error of life expectancy for the general population is approximately 0), AYLL among the HIV-diagnosed population in 25 states in 2005 was 21.1 (95% CI: 20.9 to 21.3) years, with AYLL among males 19.1 (95% CI: 18.9 to 19.3) and females 22.7 (95% CI: 22.1 to 23.3). Among males diagnosed with HIV infection in 2005, AYLL was 25.6, 18.1, and 10.2 for diagnoses at age 20, 40, and 60 years, respectively (Fig. 1). Among females diagnosed with HIV in 2005, AYLL was 30.7, 20.9, and 11.3 for diagnoses at age 20, 40, and 60 years, respectively. Whites diagnosed with HIV in 2005 had fewest AYLL compared with blacks and Hispanics at 20, 40, and 60 years [24.4, 16.9, and 9.3 (whites); 26.4, 18.1, and 10.1 (blacks); and 30.2, 23.3, and 15.3 (Hispanics)] (Fig. 2).
Based on population-based HIV surveillance data from 25 states, life expectancy after an HIV diagnosis has improved significantly (12 years) from 1996 to 2005, particularly in the first 5 years after the introduction of HAART. Significant improvements were also experienced among all race/ethnic groups and by both males (12 years) and females (11 years). Over the 10-year period, life expectancy increased 15 years for white males, 13 years for Hispanic males, and 10 years for black males. Improvements were not significant for Hispanic and white females diagnosed in the last 5 years (2001-2005). During this period, it is noteworthy that among females, the improvement in life expectancy was greatest for Hispanics. Also to be celebrated are the overall increases in life expectancy. Unfortunately, disparity in life expectancy for black and Hispanic males, compared with white males, persists. Concerning the Healthy People Objectives,4 improvements in life expectancy continue and are significant but disparities in race/ethnicity and sex continue to be of concern.
Life expectancy was worse for persons diagnosed at a later stage of disease (CD4 count <200) compared with all HIV-diagnosed persons. Late diagnoses may be due to poor access to testing or rapid progression of disease. Persons who know that they are HIV infected report fewer behavioral risk factors; therefore, prevention efforts aimed at increased testing and diagnosis should be augmented, and support for proven effective interventions should continue.24
Injection drug users had lower life expectancy consistent with their well-documented higher risk of death,12 which not only may be due to access and adherence obstacles but also may be related to prevalence of mental health, substance abuse, and comorbid conditions, including hepatitis C infection, in this population.12,25-27
Females had more AYLL than males, but, on average, females also live longer than males. AYLL was higher at younger ages due to the natural age-dependent increase in mortality. HIV-related mortality is decreasing, and this population is dying from causes similar to those of the same age-group in the general population.28,29 Hispanics had more AYLL than both blacks and whites, with the difference between Hispanics and blacks increasing at older ages. Hispanics have some of the highest rates of discontinuation of treatment and present very late to care,30 which may account for premature deaths.
Direct comparisons with previous estimates of life expectancy compared with the general population are difficult due to incompatibility of HIV populations studied [HIV diagnosed (used here) versus HIV patients in care or HIV seroconverters (used by others)]. However, a number of studies also reported the remaining excess mortality to be significant.9,12,15 Furthermore, we found that at older ages, the excess mortality persisted. For example, Bhaskaran et al12 reported that persons at least 45 years old had relatively low excess mortality, but we found that at age 60 years, AYLL was 11 years and even at age 70 years, it was still 8 years. This excess could be due to slower immune recovery after HAART initiation, late diagnosis among older age-groups, or factors related to therapy adherence.31,32 Braithwaite et al used computer modeling (simulation of patients in care) to estimate 20.4 years as the median survival, and Lohse et al10 estimated 32.5 years of survival for HIV-infected patients diagnosed and treated in HIV clinics during 2000-2005 compared with the 22.5-year life expectancy we estimated here.28
Similarly, other studies that compared an HIV-infected population with the general population reported better survival or life expectancy7-9,12,15 than we reported here, but they were all populations of HIV-infected patients in care. Our population was one of HIV-diagnosed persons, regardless of care. Estimates based on patients in care likely overestimate (more than ours do) life expectancy in the general HIV-infected population. One study by Lai et al,33 which used NCHS mortality data, looked at potential gains in life expectancy after hypothetical elimination of HIV. They found a rapid reduction from 1987 to 2000 in burden of HIV on life expectancy, especially for black males of working age.
The difference in life expectancy narrowed from 1996 to 2005 and from age 13 to age 70 years in 2005 in these 25 states. However, excess mortality (AYLL 21 years in 2005) remains in the HIV-diagnosed population compared with the US general population. In addition to their infections, HIV-infected persons may also differ from the general population on behavioral risk factors and socioeconomic factors, both of which can impact mortality and may account for some of the excess mortality reported here. HIV-infected persons have been found to smoke, consume alcohol, and inject drugs and are infected with hepatitis C at higher rates than the general population.34-39
US national HIV surveillance data routinely undergo data quality procedures related to elimination of duplicate reports, completeness and timeliness of reporting, and validity of dates of diagnosis and death. Using population-based surveillance data of the HIV-diagnosed population, we present a realistic representation of life expectancy in the HIV-diagnosed population in 25 states. To our knowledge, these estimates have not been previously provided.
These results can guide clinician communication with patients about their life expectancies after HIV diagnosis and policy makers and researchers to estimate costs and allocate resources. These estimates of life expectancy and AYLL can be added to the tools providers use to help guide discussions with patients. They also can be useful for individual planning and establishing general expectations as they help to provide a more complete picture of burden from HIV infection, along with other measures, for example, absolute or relative measures of mortality that are abundant in the literature.
Our estimates overcome some limitations of a similar cohort analysis; however, we note that HIV surveillance data are subject to several limitations. Although our data were from the largest set of population-based data from 1996 currently available for persons infected with HIV, the 25 states used in this analysis are not nationally representative. The data represent 27% of all AIDS cases diagnosed among adults and adolescents in the United States during 1996-2005. Data from states with high AIDS morbidity (eg, California and New York) were not included, which may have resulted in an over- or underrepresentation of cases among blacks and Hispanics. Blacks accounted for 49% and Hispanics 18% of AIDS cases from 1996 to 2005, whereas in the data used here from 25 states, blacks comprised 55% and Hispanics 9%.
Second, HIV surveillance data are subject to local variations in reporting and data quality. Also, HIV surveillance data do not include individuals tested anonymously or HIV-infected individuals who have not been tested. It is estimated that persons infected with HIV but not yet tested represented about 21% of the total HIV-infected population in the United States in 2006.40
Death reporting timeliness varies from state to state. Because there are delays in death reporting, data have been adjusted for death reporting delays. However, these life expectancy estimates may still be over- or underestimates of true life expectancy. Finally, these results are not adjusted for coinfection with hepatitis C virus or other comorbid conditions, or for treatment, none of which is collected by the surveillance system.
Although life expectancy and AYLL have significantly improved for persons diagnosed from 1996 to 2005 in 25 states, life expectancy remains shorter than that for the general US population and sex and race/ethnic disparities persist. Ongoing monitoring of life expectancy is important as therapies advance and other variables (eg, antiretroviral resistance, toxicities, adherence, and HIV testing practices) have the potential to change as well.41,42 AYLL can be used in conjunction with other measures of HIV disease burden, for example, mortality and survival rates, to provide policy makers, clinicians, and patients a more comprehensive picture of burden to the HIV-infected individual. As additional data become available in the national HIV surveillance system, better assessment of how well collective efforts have affected (lengthened) life expectancy for HIV-diagnosed persons will be possible.
Given the findings, further promotion of testing initiatives43 and opt-out testing will be paramount to increasing the proportion of persons who are aware of their HIV infections and who then reduce behavioral risk factors to slow the transmission of HIV. Programs aimed at increasing access to care of HIV-infected persons may improve the proportion of persons being treated and thus lengthen survival.
The authors thank the staff in 25 health departments who collected the HIV surveillance data used here.
1. Mocroft A, Vella S, Benfield TL, et al. Changing patterns of mortality across Europe in patients infected with HIV-1. Lancet
2. Palella FJ Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med
3. Adult Prevention and Treatment of Opportunistic Infections Guidelines Working Group. Guidelines for Prevention and Treatment of Opportunistic Infections in HIV-Infected Adults and Adolescents. March 24, 2009. MMWR 2009; 58 (early release) pp. 1-198. Available at: http://www.cdc.gov/mmwr/preview/mmwrhtml/rr58e324a1.htm
. Accessed December 11, 2008.
4. U.S. Department of Health and Human Services. Healthy People 2010
. 2nd ed. With Understanding and Improving Health and Objectives for Improving Health. 2 vols. Washington, DC: U.S. Government Printing Office, November 2000. Available at: http://www.healthypeople.gov/default.htm
. Accessed August 30, 2009.
5. Gardner JW, Sanborn JS. Years of potential life lost (YPLL)-what does it measure? Epidemiology
6. Burnet NG, Jefferies SJ, Benson RJ, et al. Years of life lost (YLL) from cancer is an important measure of population burden-and should be considered when allocating research funds. Br J Cancer
7. Jaggy C, von Overbeck J, Ledergerber B, et al; and the Swiss HIV Cohort Study. Mortality in the Swiss HIV Cohort Study (SHCS) and the Swiss general population. Lancet
8. Keiser O, Taffe P, Zwahlen M, et al; and the Swiss HIV Cohort Study. All cause mortality in the Swiss HIV cohort study from 1990 to 2001 in comparison with the Swiss population. AIDS
9. van Sigham A, Danner S, Ghani AC, et al; on behalf of the ATHENA National Observation Cohort Study. Mortality in patients with successful initial response to highly active antiretroviral therapy is still higher than in non-HIV-infected individuals. J Acquir Immune Defic Syndr
10. Lohse N, Hansen AB, Pedersen G, et al. Survival of persons with and without HIV infection in Denmark, 1995-2005. Ann Intern Med
11. The Antiretroviral Therapy Cohort Collaboration. Life expectancy of individuals on combination antiretroviral therapy in high-income countries: a collaborative analysis of 14 cohort studies. Lancet
12. Bhaskaran K, Hamouda O, Sannes M, et al; for the CASCADE Collaboration. Changes in the risk of death after HIV seroconversion compared with mortality in the general population. JAMA
13. Hall HI, McDavid K, Ling Q, et al. Determinants of progression to AIDS or death after HIV diagnosis, United States
, 1996 to 2001. Ann Epidemiol
14. McDavid Harrison K, Ling Q, Song R, et al. County-level socioeconomic status and survival after HIV diagnosis, United States
. Ann Epidemiol
15. Lima VD, Hogg RS, Harrigan PR, et al. Continued improvement in survival among HIV-infected individuals with newer forms of highly active antiretroviral therapy. AIDS
16. Guidelines for defining public health research and public health non-research. Centers for Disease Control and Prevention Web site. Available at: http://www.cdc.gov/od/science/regs/hrpp/researchdefinition.htm
. Revised October 4, 1999. Accessed December 7, 2008.
17. Protection of human subjects. 45 CFR Part 46. Department of Health and Human Services Web site. Available at: http://www.hhs.gov/ohrp/documents/OHRPRegulations.pdf
. Revised June 23, 2005. Accessed December 7, 2008.
18. Arias E. United States Life Tables, 2004. National Vital Statistics Reports
. Vol 56. No 9. Hyattsville, MD: National Center for Health Statistics; 2007. Available at: http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_09.pdf
. Accessed February 16, 2009.
19. Green TA. Using surveillance data to monitor trends in the AIDS epidemic. Stat Med
20. McDavid Harrison K, Kajese T, Hall IH, et al. Risk factor redistribution of the national HIV/AIDS surveillance data: an alternative approach. Public Health Rep
21. Kung HC, Hoyert DL, XU J, et al. Deaths: Final Data for 2005. National Vital Statistics Reports
. Vol 56. No 10. Hyattsville, MD: National Center for Health Statistics; 2008. Available at: http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_10.pdf
. Accessed January 10, 2009.
22. National Center for Health Statistics. Estimates of the July 1, 2000-July 1, 2007, United States
resident population from the Vintage 2007 postcensal series by year, county, age, sex, race, and Hispanic origin, prepared under a collaborative arrangement with the U.S. Census Bureau. Available at: http://www.cdc.gov/nchs/about/major/dvs/popbridge/datadoc.htm
. Accessed January 10, 2009.
23. Centers for Disease Control and Prevention. Revised surveillance case definitions for HIV infection among adults, adolescents, and children aged < 18 months and for HIV infection and AIDS among children aged 18 months to <13 years-United States
, 2008. MMWR Recomm Rep
24. Lyles CM, Kay LS, Crepaz N, et al. Best-evidence interventions: findings from a systematic review of HIV behavioral interventions for US populations at high risk, 2000-2004. Am J Public Health
25. Chander G, Himelhoch S, Moore RD. Substance abuse and psychiatric disorders in HIV-positive patients: epidemiology and impact on antiretroviral therapy. Drugs
26. Orckstroh JK, Mocroft A, Soriano V, et al. Influence of hepatitis C virus infection on HIV-1 disease progression and response to highly active antiretroviral therapy. J Infect Dis
27. Palacio H, Shiboski CH, Yelin EH, et al. Access to and utilization of primary care services among HIV-infected women. J Acquir Immune Defic Syndr
28. Braithwaite RS, Justice AC, Chang CH, et al. Estimating the proportion of patients infected with HIV who will die of comorbid diseases. Am J Med
29. Adih WK, Selik RM. Trends in Diseases Reported on Death Certificates that mentioned HIV Infection, USA, 1996-2005. Orally presented at: the APHA Annual Meeting; San Diego, CA; October 29, 2008.
30. Losina E, Schackman B, Sadownik S, et al. Disparities in survival attributable to suboptimal HIV care in the US: influence of gender and race/ethnicity. Presented at: 14th Conference on Retroviruses and Opportunistic Infections, February 27, 2007; Los Angeles, CA.
31. Viard JP, Mocroft A, Chiesi A, et al. Influence of age on CD4 cell recovery in human immunodeficiency virus-infected patients receiving highly active antiretroviral therapy: evidence from the EuroSIDA study. J Infect Dis
32. Hinkin CH, Hardy DJ, Mason KI, et al. Medication adherence in HIV-infected adults: effect of patient age, cognitive status, and substance abuse. AIDS
. 2004;18(Suppl 1):S19-S25.
33. Lai DJ, Tarwater PM, Hardy RJ. Measuring the impact of HIV/AIDS, heart disease and malignant neoplasms on life expectancy in the USA from 1987 to 2000. Public Health
34. Galvan FH, Bing EG, Fleishman JA, et al. The prevalence of alcohol consumption and heavy drinking among people with HIV in the United States
: results from the HIV Cost and Services Utilization Study. J Stud Alcohol
35. Mamary EM, Bahrs D, Martinez S. Cigarette smoking and the desire to quit among individuals living with HIV. AIDS Patient Care STDs
36. Friis-Moller N, Weber R, Reiss P, et al. Cardiovascular disease risk factors in HIV patients-association with antiretroviral therapy: results from the DAD study. AIDS
37. Centers for Disease Control and Prevention. HIV/AIDS and Viral Hepatitis. Available at: http://www.cdc.gov/hepatitis/Populations/hiv.htm
. Accessed February 22, 2009.
38. Myers T, Allman D, Xu K, et al. The prevalence and correlates of hepatitis C virus (HCV) infection and HCV-HIV co-infection in a community sample of gay and bisexual men. Int J Infect Dis
. 2009. In press.
39. Des Jarlais DC, Semaan S. HIV prevention for injecting drug users: the first 25 years and counting. Psychosom Med
40. Centers for Disease Control and Prevention. HIV prevalence estimates-United States
, 2006. MMWR Morb Mortal Wkly Rep
41. Phillips AN, Dunn D, Sabin C, et al. Long term probability of detection of HIV-1 drug resistance after starting antiretroviral therapy in routine clinical practice. AIDS
42. Zwahlen M, Lundgren JD. Commentary: death in the era of potent antiretroviral therapy: shifting causes, new challenges. Int J Epidemiol
43. Centers for Disease Control and Prevention. Advancing HIV prevention: new strategies for a changing epidemic-United States
. MMWR Morb Mortal Wkly Rep