Incidence and risk of lung cancer in HIV-infected patients

Abstract

Purpose

Lung cancer is one of the most common non-AIDS-defining malignancies among HIV-infected patients. The incidence of lung cancer has significantly increased in the HIV-positive population in recent years. The purpose of this study was to summarize the incidence and risk of lung cancer in published population-based studies of people with HIV/AIDS.

Methods

Published literature from PubMed, Embase, the Web of Science, and Google Scholar was retrieved. Sixty-five publications were selected and assessed for the following parameters: research coverage and location; continent; study period; duration of follow-up; lung cancer cases; HIV cases; incidence rate; and overall SIR or adjusted IRR. In addition, the risk of lung cancer was compared based on age, gender, HIV exposure category, CD4 count, and periods with highly active antiretroviral therapy (HAART).

Results

Lung cancer risk was greater among HIV-infected individuals compared with the general population. SIRs or adjusted IRRs were 1.5–3.4 in Europe, 0.7–6.9 in the USA, and 5.0 in Africa. Most, but not all studies did not observe a significant change in the incidence and risk of lung cancer between the pre-HAART and HAART eras. In most studies, the risk of lung cancer was higher among women, younger individuals, and injection drug users (IDUs), but the incidence of lung cancer was higher among men and the elderly. No significant trend in lung cancer risk across CD4 cell count categories was reported among the selected articles.

Conclusion

Our study suggests an increase in the incidence and risk of lung cancer in HIV/AIDS population is worldwide. The effect of HAART on the incidence and risk of lung cancer is in dispute. The risk of lung cancer based on gender differences, especially among females, as well as IDUs, requires further investigation.

Introduction

It was Irwin who reported the first case of HIV and lung cancer (1984). Since then, there have been several studies that have reported an increase in the risk of lung cancer among HIV-infected patients compared with the general population (Monfardini et al. 1989; Sridhar et al. 1992; Karp et al. 1993; Rabkin and Yellin 1994).

The introduction of highly active antiretroviral therapy (HAART) in 1996 led to a significantly prolonged survival of patients with HIV (Mocroft et al. 1998, 2002; Palella et al. 1998; Cohen et al. 2002), decreased mortality due to AIDS, and increased mortality due to malignancies (Crum-Cianflone et al. 2009). Lung cancer is one of the most common non-AIDS-defining malignancies (NADCs) among HIV-infected patients (Engels 2009; Lavole et al. 2009).

Because of the large number of cohort studies on lung cancer, we determined the incidence and risk of lung cancer in published population-based studies of people with HIV/AIDS in different continents in the pre-HAART and HAART eras among different patient groups.

Materials and methods

We conducted a comprehensive search of PubMed, Embase, the Web of Science, and Google Scholar using the following terms: lung cancer; lung carcinoma; lung neoplasm; bronchogenic carcinoma; lung neoplasms; HIV; human immunodeficiency virus; acquired immune deficiency syndrome; acquired immunodeficiency syndrome; AIDS; cancer; cancers; malignancy; and malignancies. We also retrieved three major Chinese databases, including the China Knowledge Resource Integrated (CNKI) database, the WanFang database, and the VIP database. The final literature search was conducted on May 23, 2013. The full text of the candidate articles was reviewed carefully to determine whether or not the articles met the inclusion criteria for our study. First, the article had to include the continent and region of the study, the number of HIV cases, and the number of lung cancer cases. Second, follow-up for lung cancer should start after the diagnosis of HIV was made or the onset of AIDS. Third, the article should have been published in English or Chinese.

Results

Two hundred twenty-four articles were identified using our search strategy; sixty-five publications were included. None of the articles in Chinese met the inclusion criteria. Tables 1 and 2 present a brief description of these studies in five continents (Europe, America, Australia, Asia, and Africa). Table 1 shows studies involving lung cancer in HIV-infected patients, and Table 2 shows studies involving lung cancer among patients with AIDS. To avoid the inclusion of prevalent lung cancer cases, we recorded lung cancer that occurred after the diagnosis of HIV was made or the onset of AIDS in the selected studies, and excluded articles that did not provide data during these periods. Several articles did not demonstrate the observation starting time before or after the diagnosis of HIV was made or the onset of AIDS, and we did not exclude those articles. Most researchers estimated the risk of lung cancer using a standardized incidence ratio (SIR) and comparing the number of observed cases to the expected number of cases estimated from the general population. The adjusted incidence rate ratio (IRR), adjusted for age, gender, race/ethnicity, smoking, and route of infection was also calculated.

Table 1.

Characteristics of studies on lung cancer in HIV-infected patients

First author (year of publication)	Research coverage and location/continent	Study period/duration of follow-up	Lung cancer/ HIV cases	Incidence rate (per 100,000 person-years)	Overall SIR or adjusted IRR (95 % CI)
Ferraresi et al. (2012)	Brescia, Italy/Europe	1999–2009	23/5,096		3.4 (2.26–5.12)
Calabresi et al. (2013)	Brescia, Italy/Europe	1999–2009	23/5,090		3.4 (2.3–5.1)
Franzetti et al. (2012)	Authors’ institution in Milan, Italy/Europe	1985–2011/at least 6 months	22/5,924
Bower et al. (2003)	Authors’ hospital in London, UK/Europe	1986–2001	11/8,640
Engsig et al. (2011)	Denmark, nationwide/Europe	1995–2009	29/5,053		2.38 (1.61–3.53)
Serraino et al. (2000)	Men, Southeast France, Italy/Europe	1982–1998/2 months after enrollment	4/5,281		2.9 (0.7–7.4)
Santos Gonzalez et al. (1996)	Authors’ institution in Málaga, Spain/Europe	1991–1995	7/1,258
Estrada et al. (1996)	Authors’ center in Barcelona, Spain/Europe		5/2,586
Vogel et al. (2011)	Bonn, Germany/Europe	1996–2009	5/1,476
Serraino et al. (2007)	Southeast France, Italy/Europe	1988–2004	14/8,074		1.7 (0.9–2.8)
Franceschi et al. (2010)	7 large hospitals in different Swiss cities/Europe	1985–2006/3 months after enrollment	30/9,429
Powles et al. (2009)	London, UK/Europe	1983–2007	18/11,112		1.95 (1.16–3.09)
Clifford et al. (2005)	7 large hospitals in different Swiss cities/Europe	1985–2002/3 months after enrollment	14/7,304		3.2 (1.7–5.4)
Herida et al. (2003)	61 university hospitals in France/Europe	1992–1999/mean, 32 months	112/77,025
Petruckevitch et al. (1999)	African in London, 11 units in UK/Europe	1985–1995/mean, 20.7 months	2/1,056
Guiguet et al. (2009)	62 university hospitals in France/Europe	1998–2006/mean 4.9 years	207/52,278	85
Bedimo et al. (2004)	Birmingham, UK/Europe	1989–2002	3/2,882
Spagnuolo et al. (2012)	Authors’ clinic in Milan, Italy/Europe	1991–2010	24/6,495	34
Dauby et al. (2011)	Brussels, Belgium/Europe	2002–2009/mean, 49 months	5/3,126
Prosperi et al. (2010)	Italy/Europe	1991–2002/mean, 58 months	9/6,695
Bruyand et al. (2009)	Southwest France/Europe	1998–2006	30/4,194
Johnson et al. (1997)	6 regions in USA/N. America	1988–1994	5/1,073
Cooksley et al. (1999)	Harris County, Texas, USA/N. America	1985–1994	18/14,986		0.7 (0.4–1.1)
Patel et al. (2008)	13 regions in USA/N. America	1992–2003	140/54,780	88.8	3.3 (2.8–3.9)
Levine et al. (2010)	Women, 6 regions in USA/N. America	1994–2006/mean, 5.8 years	12/2,651	63.7	3.28 (1.70–5.74)
Sigel et al. (2012)	Veterans, USA/N. America	1997–2008/6 months after registration/mean, 5.8 years	457/37,294	204	1.7 (1.5–1.9)
Bedimo et al. (2009)	Veterans, USA/N. America	1997–2004/mean, 5.1 years	504/33,420	287.9	2.0 (1.8–2.2)
Shiels et al. (2010a)	Injection drug users, Baltimore, USA/N. America	1988–2006/at least 2 years	13/1,072
Ruiz (2010)	Authors’ institution in Louisiana, USA/N. America	2002–2009	16/2,060
Engels et al. (2008)	3 regions in USA/N. America	1991–2002/4–60 months after registration	109/57,350	59	2.6 (2.1–3.1)
Engels et al. (2006a)	Moore Clinic in Baltimore, USA/N. America	1989–2003	33/5,238	170	6.9 (4.8–9.7)
Phelps et al. (2001)	Women, 4 sites in USA/N. America	1993–2000	4/871	95	6.39 (3.71–11.02)
Ricaurte et al. (2001)	Authors’ institution in New York, USA/N. America	1998/1 year	5/2,616	191.1	3.01 (1.3–7.0)
Albu et al. (2000)	Authors’ hospital in New York, USA/N. America	1993–1998	17/3,578
Parker et al. (1998)	Texas, USA/N. America	1990–1995	36/26,181		6.5 (4.5–8.9)
Krishnan et al. (2011)	USA/N. America	1998–2008/mean, 3.8 years	6/3,158
Seaberg et al. (2010)	Homosexual men, 4 regions in USA/N. America	1984–2007/mean, 9.7 years	9/3,526		1.09 (0.50–2.07)
Silverberg et al. (2011)	California, USA/N. America	1996–2008	56/20,775	62	1.2 (0.9–1.6)
Silverberg et al. (2009)	California, USA/N. America	1996–2007/mean, 4.2 years	54/20,277	63.6
Burgi et al. (2005)	Multicenter, USA/N. America	1988–2003	3/4,144
Long et al. (2008)	An urban HIV clinic in Baltimore, USA/N. America	1996–2005/mean, 4.5 years	29/2,566		5.5 (3.7–8.0)
Pinto Neto et al. (2012)	An HIV clinic in Vitória, Brazil/S. America	2010–2011	3/730
van Leeuwen et al. (2009)	Australia, nationwide/Australia	1982–2004/mean, 7.9 years	37/20,232
Petoumenos et al. (2013)	27 sites in Australia/Australia	1999–2005	6/2,181
Kiertiburanakul et al. (2007)	Ramathibodi Hospital in Thailand/Asia	1999–2003	2/1,416
Oh et al. (1999)	Seoul, South Korea/Asia	1985–1998	1/173
Mbulaiteye et al. (2006)	Uganda/Africa	1988–2002	3/12,607		5.0 (1.0–15)
Lifson et al. (2010)	318 sites in 33 countries/5 continents	2002–2006/mean, 33 months	18/5,472
Silverberg et al. (2007)	318 sites in 33 countries/5 continents	2002–2006	8/5,472
Borges et al. (2013)	Control arms of 3 randomized trials/5 continents	Mean, 59 months	26/5,023
Srisawat et al. (2008)	Netherlands, Australia, Thailand/3 continents	1997–2007/mean, 6.3 years	2/863		27.32

Table 2.

Characteristics of studies on lung cancer in patients with AIDS

First author (year of publication)	Research coverage and location/continent	Study period/duration of follow-up	Lung cancer/ HIV cases	Incidence rate (per 100,000 person-years)	Overall SIR or adjusted IRR (95 % CI)
Dal Maso et al. (2003)	Multicenter, Italy/Europe	1985–1998/4–42 months after AIDS onset	8/12,104		5.4 (2.3–10.8)
Shebl et al. (2010)	11 regions in USA/N. America	1977–2002/4–120 months after AIDS onset	853/322,675	82.6
Shiels et al. (2010b)	15 regions in USA/N. America	1996–2007/4–60 months after AIDS onset	605/212,055		3.0 (2.8–3.2)
Simard et al. (2010)	15 regions in USA/N. America	1980–2004/3–10 years after AIDS onset	888/263,254
Chaturvedi et al. (2007)	11 regions in USA/N. America	1980–2002/4–27 months after AIDS onset	393/317,007	82.3	2.9 (2.6–3.2)
Engels et al. (2006b)	11 regions in USA/N. America	1980–2002/4–27 months after AIDS onset	393/375,933
Simard et al. (2012)	Childhood, 15 regions in USA/N. America	1980–2007/4–120 months after AIDS onset	1/5,850
Ramirez-Marrero et al. (2010)	Hispanics, Puerto Rico, USA/N. America	1987–2003/began 3 months after AIDS onset	21/28,460
Biggar et al. (2005)	New York, USA/N. America	1980–2000	361/113,877
Frisch et al. (2001)	11 regions in USA/N. America	1978–1996/4–27 months after AIDS onset	241/302,834		2.8 (2.4–3.1)
Gallagher et al. (2001)	New York, USA/N. America	1981–1994/3–60 months after AIDS onset	101/122,993
Mbulaiteye et al. (2003)	11 regions in USA/N. America	1990–1996/4–27 months after AIDS onset	74/147,596		2.8
Simard et al. (2011)	15 regions in USA/N. America	1980–2006/0–5 years after AIDS onset	1698/472,378
Grulich et al. (2002)	Australia, nationwide/Australia	1985–1999/6–24 months after AIDS onset	4/8,118		4.22
Grulich et al. (1999)	New South Wales, Australia/Australia	1980–1993/began 6 months after AIDS onset	2/3,616		3.88

To be consistent in follow-up time, we just list lung cancer cases after AIDS onset

Nearly all studies demonstrated that the risk of lung cancer was greater among HIV-infected patients compared with the general population. In Europe, the SIR or adjusted IRR for lung cancer in HIV-infected patients ranged from 1.5 to 3.4. With respect to AIDS, we identified one multicentric study that included 12,104 persons with AIDS and 42 months of follow-up in Italy; the SIR was up to 5.4 (Dal Maso et al. 2003). In the USA, the SIR or adjusted IRR of lung cancer in HIV-infected patients ranged from 0.7 to 6.9, and the SIR had a 2.8–3.0-fold increase in patients with AIDS. Studies in Asia, Africa, and Australia meeting our criteria were limited, and only one study in Uganda reported a SIR of 5.0, suggesting a 5.0-fold increase in the risk of lung cancer in the HIV population compared with the general Ugandan population (Mbulaiteye et al. 2006).

It is difficult to compare the incidence of lung cancer in the HIV or AIDS population in different countries or continents with the aforementioned studies because of different durations of follow-up, enrollment criteria, and survey designs. We identified one study involving 318 sites in 33 countries between 2002 and 2006, but two articles from the same study provided different lung cancer cases and neither provided SIR data. We chose multicentric, large sample cohort studies to further identify the risk of lung cancer in the HIV populations in different countries. The Danish HIV Cohort study was a population-based prospective nationwide cohort study of all HIV patients ≥16 years of age at the time of diagnosis and who were treated at Danish HIV centers after January 1, 1995. The study included all mothers and fathers of the HIV patients and population controls. The researchers found that HIV patients had an increased risk of lung cancer (adjusted IRR 2.38), and both fathers and mothers of HIV patients had an increased risk of lung cancer (adjusted IRR for fathers 1.31; adjusted IRR for mothers 1.35; (Engsig et al. 2011)). Patel et al. analyzed data from two large prospective cohort studies that included 54,780 HIV patients in 13 geographic areas in the USA. They compared the incidence rate of lung cancer among these persons with the general population between 1992 and 2003 and found 140 lung cancer cases (SIR 3.3). Patel et al. also determined the SIR for three time periods [pre-HAART (SIR 3.5), early-HAART (SIR 3.8), and late-HAART (SIR 3.6); 2008]. Several studies compared the incidence or risk of lung cancer in the HIV population by periods in relation to HAART and AIDS onset, as well as age, gender, HIV exposure category, and CD4 count (Tables 3, 4, 5, 6, 7, 8, 9).

Table 3.

SIRs and 95 % CIs for lung cancer in patients with HIV/AIDS in the pre-HAART and HAART eras

First author (year of publication)	Country	HIV or AIDS	Pre-HAART era	HAART era
				Early-HAART	Late-HAART
Franceschi et al. (2010)	Swiss	HIV	3.3 (1.4–6.6)	2.8 (1.3–5.1)	2.6 (1.3–4.6)
Powles et al. (2009)	UK	HIV	0.00 (0.00–1.52)	3.10 (1.34–6.11)	2.37 (1.14–4.36)
Herida et al. (2003)	France	HIV	Men 1.13 (0.71–1.72) Women 1.08 (0.01–5.98)	Men 2.12 (1.67–2.65) Women 6.59 (3.40–11.52)
Patel et al. (2008)	USA	HIV	3.5 (2.5–4.9)	3.8 (2.8–5.0)	3.6 (2.8–4.6)
Levine et al. (2010)	USA	HIV	Women 4.97 (0.60–17.95)	Women 3.08 (1.47–5.66)
Engels et al. (2008)	USA	HIV	2.6 (1.6–4.1)	2.6 (2.1–3.2)
Engels et al. (2006a)	USA	HIV	1.7 (0.2–6.3)	5.2 (2.8–8.7)	5.3 (3.1–8.4)
van Leeuwen et al. (2009)	Australia	HIV	1.47 (0.80–2.46)	1.24 (0.54–2.45)	1.10 (0.62–1.82)
Simard et al. (2010)	USA	AIDS		2.6 (2.4–2.8)
Chaturvedi et al. (2007)	USA	AIDS	1980–1989 2.5 (1.8–3.3) 1990–1995 3.3 (2.9–3.8)	2.5 (2.1–3.1)
Engels et al. (2006b)	USA	AIDS	1980–1989 2.5 (1.9–3.3) 1990–1995 3.3 (2.9–3.8)	2.6 (2.1–3.1)
Ramirez-Marrero et al. (2010)	USA	AIDS	1987–1989 3.7 (0.1–20.7) 1990–1995 12.9 (7.4–21.0)	3.1 (0.9–8.0)

Table 4.

SIRs and 95 % CIs for lung cancer in patients with HIV by HAART use

First author (year of publication)	Country	HAART use
		No		Yes
		Cases	SIR (95 % CI)	Cases	SIR (95 % CI)
Engels et al. (2006a)	USA	14	4.2 (2.2–7.4)	19	5.7 (3.3–9.3)
Clifford et al. (2005)	Swiss	9	3.5 (1.6–6.6)	5	2.8 (0.9–6.5)
Serraino et al. (2007)	Italy	7	1.3 (0.5–2.7)	7	2.4 (1.0–5.0)

Table 5.

Incidence rate (per 100,000 person-years) for lung cancer in patients with HIV/AIDS in pre-HAART and HAART eras

First author (year of publication)	Country	HIV or AIDS	Pre-HAART era	HAART era
				Early-HAART	Late-HAART
Spagnuolo et al. (2012)	Italy	HIV	34		27
Levine et al. (2010)	USA	HIV	Women 62.1	Women 64.1
Engels et al. (2008)	USA	HIV	42	64
Engels et al. (2006a)	USA	HIV	60	180	220

Table 6.

SIRs and 95 % CIs for lung cancer in patients with HIV/AIDS by age, gender, and HIV exposure category

First author (year of publication)	Country	HIV or AIDS	Age (years)		Gender		HIV exposure category
Herida et al. (2003)	France	HIV					Pre-HAART era
							Homosexual men	0.92 (0.42–1.75)
							IDU men	3.16 (1.15–6.87)
							Heterosexual men	0.99 (0.27–2.53)
							HAART era
							Homosexual men	1.78 (1.22–2.49)
							IDU men	4.70 (2.83–7.34)
							Heterosexual men	1.92 (1.14–3.03)
Levine et al. (2010)	USA	HIV			Women	3.28 (1.70–5.74)
Calabresi et al. (2013)	Italy	HIV			Men Women	3.6 (2.4–5.4) 1.6 (0.2–11.1)
Franzetti et al. (2012)	Italy	HIV			Men Women	2.1 (1.3–3.4) 4.8 (1.3–12.2)
Engels et al. (2006a)	USA	HIV	<40 40–49 50+	8.7 (2.4–22) 6.6 (3.8–11) 3.0 (1.5–5.2)	Men Women	3.8 (2.4–5.7) 8.8 (4.4–16)	IDU Non-IDU	5.3 (3.2–8.3) 4.0 (2.2–6.7)
Serraino et al. (2000)	Italy	HIV					IDU men Homosexual men	6.2 (1.3–18.2) 1.4 (0.0–8.1)
Vogel et al. (2011)	Germany	HIV			Men	1.5 (0.5–3.1)
Clifford et al. (2005)	Switzerland	HIV			Men Women	3.0 (1.6–5.3) 4.8 (0.5–17.6)	Homo-/bisexual men IDU Heterosexual/other	1.2 (0.2–3.5) 12.6 (5.0–26.1) 3.0 (0.8–7.9)
Cooksley et al. (1999)	USA	HIV			Men Women	0.6 (0.3–0.9) 1.8 (0.4–5.1)
Seaberg et al. (2010)	USA	HIV					Homosexual	1.09 (0.50–2.07)
Shiels et al. (2010b)	USA	AIDS	0–29 30–39 40–49 50–59 60–69 70+	21.6 (7.0–50.4) 7.1 (5.4–9.0) 3.9 (3.4–4.4) 2.8 (2.5–3.2) 1.8 (1.4–2.2) 1.2 (0.66–1.9)
Chaturvedi et al. (2007)	USA	AIDS	15–29 30–39 40–49 50–59 60–69 70+	10.4 (3.8–22.7) 6.3 (5.0–7.9) 3.7 (3.1–4.3) 2.3 (1.8–2.8) 1.6 (1.2–2.2) 0.8 (0.3–1.8)	Men Women	2.8 (2.5–3.2) 3.7 (2.8–4.8)	Homosexual men IDU Homosexual men and IDU Heterosexual Other/unknown	2.6 (2.2–3.1) 3.9 (3.3–4.7) 2.4 (1.2–4.4) 2.8 (2.0–3.7) 2.5 (1.9–3.2)
Ramirez-Marrero et al. (2010)	USA	AIDS			Men
					1987–1989	2.9 (0.1–16.2)
					1990–1995	8.9 (4.9–15.0)
					1996–2003	1.4 (0.2–5.2)
					Women
					1990–1995	16.7 (2.0–60.4)
					1996–2003	10.2 (1.2–36.7)

Table 7.

Incidence rates (per 100,000 person-years) for lung cancer in patients with HIV/AIDS by age, gender, and HIV exposure category

First author (year of publication)	Country	HIV or AIDS	Age (years)		Gender		HIV exposure category
Engels et al. (2006a)	USA	HIV	<40 40–49 50+	40 240 560	Men Women	170 170	IDU Non-IDU	200 150
Chaturvedi et al. (2007)	USA	AIDS	15–29 30–39 40–49 5–59 6–69 70+	7.7 36.1 118.8 288.8 479.2 312.0	Men Women	87.3 61.7	Homosexual men IDU Homosexual men and IDU Heterosexual Other/unknown	65.9 103.4 44.0 84.3 112.2
Cooksley et al. (1999)	USA	HIV		Men	Women
			0–10	0.8	0
			10–19	0.5	0.5
			20–29	5.9	2.8
			30–39	32.6	25.9
			40–49	276.9	197.1
			50+	2,507.0	1,367.5

Table 8.

SIRs and 95 % CIs for lung cancer in patients with HIV/AIDS by CD4 count

First author (year of publication)	Country	HIV or AIDS	At AIDS onset (cells/μl)		At HIV registration (cells/μl)
Engels et al. (2008)	USA	HIV			750+ 500–749 350–499 200–349	2.2 3.2 3.1 2.9
Chaturvedi et al. (2007)	USA	AIDS	300+ 250–299 200–249 150–199 100–149 50–99 0–49	1.4 (0.6–2.8) 2.4 (1.0–4.7) 2.3 (1.1–4.2) 3.1 (2.3–3.9) 3.1 (2.3–4.1) 3.0 (2.2–4.1) 2.9 (2.3–3.8)
Mbulaiteye et al. (2003)	USA	AIDS	200+ 100–199 50–99 0–49	2.6 2.7 3.8 2.6

Table 9.

SIRs and 95 % CIs for lung cancer in patients with HIV/AIDS before and after AIDS

First author (year of publication)	Country		Before AIDS			After AIDS
			Cases	SIR (95 % CI)		Cases	SIR (95 % CI)
Clifford et al. (2005)	Switzerland		10	3.2 (1.5–6.0)		4	3.1 (0.8–8.0)
Engels et al. (2008)	USA			2.3 (1.8–2.8)			4.7 (3.0–6.9)
Dal Maso et al. (2003)	Italy	60–25 months before 24–7 months before Within 9 months	1 0 13	0.3 (0.0–1.5) 0 10.3 (5.5–17.7)	4–42 months after	8	5.4 (2.3–10.8)
Grulich et al. (2002)	Australia	5 plus years before, or never developed AIDS	4	0.95	6 months–2 years after	4	4.22
		5 years–6 months before	2	0.37
		within 6 months	7	5.42
Li et al. (2002)	Australia	5 years prior to AIDS	1	3.16 (0.45–22.45)

There was heterogeneity in the effect of HAART on risk or incidence of lung cancer in the HIV/AIDS population (Tables 3, 4, 5). We found three articles, which suggested that the incidence or risk of lung cancer in the HAART era changed compared with the pre-HAART era. Engels et al. used population-based registry data to determine the cancer incidence among 375,933 patients with AIDS in six US states and five metropolitan areas between 1980 and 2002, and found that following an apparent increase during the 1980s, the lung cancer risk declined from the pre-HAART era (1990–1995) to the HAART era (1996–2002; SIR 3.3 and 2.6, respectively; P = 0.02; 2006b). In another study, Engels et al. reported that the incidence of lung cancer increased across three periods (1989–1994, 1995–1999, and 2000–2003; P = 0.02), which corresponded to the pre-HAART, early-HAART, and late-HAART eras, respectively (2006a). Ramírez-Marrero et al. observed that the SIR of lung cancer in Hispanics patients with AIDS in Puerto Rico was 12.9 in the pre-HAART era and 3.1 in the early-HAART era, respectively, compared with the general population (2010). Of note, most studies did not observe a significant change in the risk or incidence of lung cancer between the pre-HAART and HAART eras, and none of the three studies which compared lung cancer risk among HIV patients using HAART or not identified a significant change in the SIR (Table 4).

In contrast to the incidence rate trends (Table 7), the SIRs of lung cancer were higher among women and younger individuals in most studies (Table 6). Engels et al. noted that even though the incidence of lung cancer was similar in males and females, the SIR was twofold higher in women than men (P = 0.02), and even though the incidence rates increased with age, the SIRs actually declined with age (P = 0.02; 2006a). Also, Shiels reported that the SIRs for lung cancer among people with AIDS were highest in the youngest ages and declined significantly across age groups (P trend < 0.001; 2010b). Similarly, Cooksley and colleagues reported that between 1985 and 1994 in Harris County, Texas, the risk of lung cancer was threefold higher in women than men (n.b., the authors did not provide a P value for the comparison of SIRs; 1999). Most articles listed in Table 6 suggest that the SIRs of lung cancer are higher in injection drug users (IDUs) than other HIV transmission groups. Clifford reported that the SIR of lung cancer in IDUs was up to 12.6, but the SIRs were only 1.2 and 3.0 in homo-/bisexual men and heterosexual/other patients with HIV in Switzerland, respectively (2005).

No significant trend in lung cancer risk across CD4 cell count categories was reported among the selected articles (Table 8). Some studies suggested that the risk of lung cancer after AIDS onset increased compared with before AIDS onset (Table 9). Grulich noted that the SIR of lung cancer was elevated within 6 months before or after the onset of AIDS compared with 5 plus years before, 5 years to 6 months before, and 6 months to 2 years after AIDS onset (P < 0.001; 2002). Dal Maso reported similar results (n.b., the author did not show the P value; 2003), and Engels noted a twofold increased risk of lung cancer after the diagnosis of AIDS than before the diagnosis of AIDS (SIR = 4.7 and 2.3, respectively; (2008).

Discussion

Our survey highlights several important findings on lung cancer and HIV. First, in most countries in the world, the risk of lung cancer in the HIV population is significantly higher than the general population. Second, the effect of HAART on the incidence and risk of lung cancer is in dispute. Third, although the lung cancer incidence rates are lower in females and young people, the risks of lung cancer are higher compared with males and the elderly. Fourth, IDUs are at a higher risk of lung cancer than other HIV transmission populations.

Lung cancer is the most frequent NADCs in HIV-infected patients (Engels 2009; Lavole et al. 2009). The highest reported SIR for lung cancer is 27.32, which was calculated in a retrospective study at the HIV-Netherlands/Australia/Thailand Research Collaboration between 1997 and 2007 (Srisawat et al. 2008). Two lung cancer patients in 863 subjects were identified; thus, the high SIR may be due to the small number of cases. Although little or no increase has been observed in several studies (Cooksley et al. 1999; Seaberg et al. 2010), a significant increase in the incidence and risk of lung cancer among HIV-infected patients has been reported by most researchers. One explanation for this increase may be the immunosuppression of the HIV population, which is often identified based on the CD4 count, but studies focusing on the relationship between the SIR and CD4 count at the onset of AIDS showed negative results (Mbulaiteye et al. 2003; Chaturvedi et al. 2007). Some authors have attributed the increase in the incidence and risk of lung cancer to the excessive cigarette smoking in HIV-positive patients. Ferraresi reported on 23 HIV patients with lung cancer between 1999 and 2009 in author’s clinic; all subjects except two were current smokers (2012). All HIV patients diagnosed with lung cancer were current or former smokers from the Danish HIV Cohort Study by Engsig et al. (2011). Current and former smokers had a substantially greater risk of lung cancer compared with never smokers (current smokers IRR 6.3; former smokers IRR 3.0; Sigel et al. 2012). After adjusting for smoking status, even though the SIR was approximately one-half of the unadjusted SIR, HIV infection was still associated with an increased risk of lung cancer (Engels et al. 2006a), thus smoking did not account for all of the lung cancer risk in the HIV population. HIV patients with coexisting COPD, bacterial pneumonia, or asthma have a higher incidence rate of lung cancer (Kirk et al. 2007); thus, Sigel et al. adjusted for age, gender, race/ethnicity, smoking prevalence, previous bacterial pneumonia, and COPD, and showed that the adjusted IRR of lung cancer associated with HIV infection remained significant (IRR 1.7; 2012). Sigel et al. concluded that HIV infection is an independent risk factor for lung cancer.

The widespread use of HAART has improved the survival of HIV-infected patients (Palella et al. 1998) and has led to a decline in the incidence of AIDS-defining malignancies, such as Kaposi sarcoma (Besson et al. 2001). As a result, people are gradually paying attention to the effect of HAART on NADCs. Several studies have demonstrated an increased incidence and broader spectrum of NADCs during the HAART era (Hessol et al. 2007; Long et al. 2008; Crum-Cianflone et al. 2009; Franzetti et al. 2012). Lung cancer is one of the most common malignancies among NADCs; thus, it is plausible that HAART itself could affect the incidence and risk of lung cancer; however, most researchers have not demonstrated a significant change in the incidence or risk of lung cancer in HAART era or after HAART use. We found three articles to support the notion that the incidence or risk of lung cancer in the HAART era has changed compared with the pre-HAART era. Even though Engels reported that the incidence of lung cancer increased in the HAART era, when the incidence of lung cancer was compared with HAART use, there was no statistical difference (P = 0.1; 2006a). Moreover, Engels did not find a statistical significance difference in the SIRs of lung cancer in the HIV population during the HAART era or following HAART use (2006a). Engels also reported that the risk of lung cancer in the AIDS population declined from the pre-HAART era to the HAART era (SIR 3.3 and 2.6, respectively; P = 0.02; 2006b); of note, the risk of lung cancer was not compared to HAART use. We speculate that two differences lead to these disparities. First, population is different, a HIV-infected population, or an AIDS population. Second, the duration of follow-up in the AIDS population was 4–27 months after AIDS onset, but in the HIV population, it is impossible to determine the actual duration of follow-up, so that in our inclusion criteria, follow-up for lung cancer should start after the diagnosis of HIV was made or the onset of AIDS; specifically, we labeled the HIV or AIDS population as well as the duration of follow-up in all tables, so it was easier for readers to compare and understand the data among different studies. The third article was written by Ramírez-Marrero et al. who estimated the risks of AIDS-defining cancers and NADCs among Hispanics with AIDS in a registry linkage study that included 28,460 subjects between 1987 and 2003 in Puerto Rico, and found that the SIR of lung cancer declined fourfold in the HAART era compared with the pre-HAART era (2010). We speculate that the results may be due to different races (Hispanics in this study, other than White and Black in most other studies), as well as the mode of common HIV transmission, IDUs in Puerto Rico, and other than homosexuals in the USA. Moreover, the small number of cases for lung cancer between 1987 and 1989 (only 1) may affect the stability of the SIR, thus additional studies focusing on the incidence and risk of lung cancer in the HIV or AIDS population, as well as an analysis of risk factors, such as smoking, gender, age, behavior, and a family history of lung cancer, should be conducted among Hispanics.

Most articles have shown that the incidence of lung cancer in the HIV population is higher for males, but the risk is higher for females, which was consistent with the world trends of lung cancer regardless of HIV infection in the general population. Indeed, the total lung cancer incidence rates have been declining among males in many areas of the world, but rising rapidly among females (Devesa et al. 2005; Fu et al. 2005). How to explain this phenomenon is an intense research interest. Some authors have demonstrated that the predisposition to lung cancer in females is in association with two genes, CYP1A1 and GSTM1 (Ryberg et al. 1994). In addition, exposure to cooking may also increase the risk of lung cancer in women (Chiu et al. 2010; Lee and Gany 2012). Whether or not estrogen and receptors present in lung cancer cells play a role is unclear. Some case–control studies have suggested that women are more susceptible to tobacco carcinogens than men (Brownson et al. 1992; Harris et al. 1993; Risch et al. 1993), but most cohort studies have shown no gender differences in lung cancer susceptibility to smoking exposure (Prescott et al. 1998; Bain et al. 2004; Jee et al. 2004; Freedman et al. 2008). Matteis used data from a large population-based case–control study in Italy, which contained detailed information on lifetime smoking habits, direct interviews with cases, and a high response rate in both cases and controls, and found a clear negative interaction between female lung cancer and tobacco smoking (2013). With respect to HIV-infected women, smoking history and pack-years of smoking are still the most important risk factors for lung cancer. Among HIV-uninfected women, most lung cancers are observed in women with at least 20 pack-years of smoking, while the majority of lung cancers in HIV-infected women occur in those with histories of 10–20 pack-years of smoking (Levine et al. 2010). We have not identified a cohort or case–control study involving the relationship between smoking exposure and gender differences in lung cancer in the HIV/AIDS population; thus, the mechanism underlying the susceptibility of females to lung cancer is unclear. Not all studies have demonstrated a higher risk of lung cancer in females than males. Calabresi reported twice the risk of lung cancer in men than women (2013), while Clifford did not detect a difference in risk between men and women (2005). Interestingly, both studies were conducted in Europe, while all studies in the USA suggested a higher risk of lung cancer in the female HIV/AIDS population than the male HIV/AIDS population (Chaturvedi et al. 2007; Cooksley et al. 1999; Engels et al. 2006a; Ramirez-Marrero et al. 2010). Such regional differences deserve further investigation.

Although the incidence of lung cancer increases with age (P < 0.0001; (Cooksley et al. 1999; Engels et al. 2006a; Chaturvedi et al. 2007), the SIRs for lung cancer were highest in the youngest patients and decreased significantly across age groups (P trend < 0.001; (Shiels et al. 2010b). The SIR is defined as the ratio of observed lung cancers in the HIV/AIDS population to the number expected based on age-, gender-, race-, and calendar-year-specific lung cancer incidence rates in the general population. The incidence of lung cancer in young patients is very low, which may cause a significant increase in the SIR of lung cancer in younger patients.

The risks of lung cancer as well as liver cancer are significantly higher among IDUs than other HIV exposure groups, but the risks of Kaposi sarcoma and non-Hodgkin lymphoma are higher among homo-/bisexual men than IDUs (Serraino et al. 2000; Clifford et al. 2005). Those results suggest that different lifestyles may affect the development of cancer in HIV patients. Several studies have demonstrated that IDU is a risk factor for the development of COPD (Crothers et al. 2006; Overland et al. 1980; Sherman et al. 1987), and some have attributed the risk to tobacco use and low socioeconomic status in this population (Burns et al. 1996), which may also cause lung cancer in our opinion. In addition, most IDUs have drug inhalation histories, which would cause squamous metaplasia of the tracheobronchial epithelium and the development of lung cancer (Benson and Bentley 1995). There is a need for more epidemiologic surveys of lung cancer among IDUs with HIV.

In conclusion, the incidence and risk of lung cancer is increased in the HIV population worldwide. The effect of HAART on the incidence and risk of lung cancer is in dispute, and the risk of lung cancer based on gender differences, especially among females, as well as IDUs, warrants further investigation.