HIV Transmission Patterns Among The Netherlands,Suriname,and The Netherlands Antilles: A Molecular Epidemiological Study

Abstract

We aimed to study patterns of HIV transmission among Suriname, The Netherlands Antilles, and The Netherlands. Fragments of env, gag, and pol genes of 55 HIV-infected Surinamese, Antillean, and Dutch heterosexuals living in The Netherlands and 72 HIV-infected heterosexuals living in Suriname and the Antilles were amplified and sequenced. We included 145 pol sequences of HIV-infected Surinamese, Antillean, and Dutch heterosexuals living in The Netherlands from an observational cohort. All sequences were phylogenetically analyzed by neighbor-joining. Additionally, HIV-1 mobility among ethnic groups was estimated. A phylogenetic tree of all pol sequences showed two Surinamese and three Antillean clusters of related strains, but no clustering between ethnic groups. Clusters included sequences of individuals living in Suriname and the Antilles as well as those who have migrated to The Netherlands. Similar clustering patterns were observed in env and gag. Analysis of HIV mobility among ethnic groups showed significantly lower migration between groups than expected under the hypothesis of panmixis, apart from higher HIV migration between Antilleans in The Netherlands and all other groups. Our study shows that HIV transmission mainly occurs within the ethnic group. This suggests that cultural factors could have a larger impact on HIV mobility than geographic distance.

Introduction

U ntil 2000, male-to-male sexual contact was the most frequently reported transmission route for HIV infection in The Netherlands. Thereafter, the number of new HIV diagnoses in heterosexuals has increased, approaching the level of men having sex with men. Many of the patients who acquired HIV by heterosexual contact are of non-Dutch origin. Of these, the largest group is composed of sub-Saharan Africans (43%), followed by individuals originating from Latin America and the Caribbean (15%), predominantly Suriname and The Netherlands Antilles (further: Antilles), two former Dutch colonies.¹

In the Caribbean, heterosexual contact is the major HIV transmission route and, as in The Netherlands, the epidemic is dominated by subtype B.² The estimated HIV prevalence in the Caribbean [1.1%, 95% confidence interval (CI) 1.0–1.2%] and in Suriname (2.4%, 95% CI 1.4–2.3%) is relatively high compared to The Netherlands (0.2%, 95% CI 0.1–0.3%).^3,4 Suriname and the Antilles have an estimated population of 500,00 and 290,000, respectively; in contrast 450,000 persons of Surinamese and Antillean ethnic origin are living in The Netherlands. Studies in The Netherlands show that Surinamese and Antillean migrants frequently visit their country of origin and report unprotected sexual contacts during these visits.^5
–7 Therefore, sexual contact in the country of origin could be an important risk factor for the transmission of HIV.³ In addition, many of the HIV-infected persons of Surinamese (31%) and Antillean ethnic origin (51%) reported their country of origin as the most likely country of infection.¹

Molecular epidemiological research in The Netherlands suggests that the rise in the proportion of heterosexual HIV cases might partly be explained by immigration.⁸ Since phylogenetic analysis has become a valuable tool in studying the epidemiology of infectious diseases and reconstructing epidemiological relationships between infected persons, we used it to compare HIV strains circulating in Suriname and the Antilles with strains circulating among these ethnic groups in The Netherlands to gain more insight into epidemiological links between populations.

This study aimed to investigate whether there is evidence for HIV transmission among persons in Suriname, the Antilles, and Surinamese and Antillean persons living in The Netherlands. As disassortive ethnic mixing (i.e., partners of different ethnic origin) is not uncommon,^9,10 we also investigated whether there was evidence for heterosexual transmission of HIV between the Surinamese, the Antillean, and the indigenous Dutch populations. Finally, we provided additional analyses to estimate HIV-1 mobility between Suriname, the Antilles, and The Netherlands by using phylogenetic analysis and an algorithm that estimates the relationship among HIV strains in geographically structured populations.^11,12

Materials and Methods

Study population and sample collection

Stored serum samples taken from 127 HIV-infected individuals were collected anonymously in Suriname, the Antilles, and The Netherlands. Of these, 72 serum samples were randomly collected from recently diagnosed HIV-infected heterosexual persons living in Suriname (Dermatology Service, Paramaribo, n = 35) and the Antilles (Red Cross Blood Bank Foundation, Curaçao, n = 37). The remaining 55 samples, randomly collected at various HIV treatment centers in The Netherlands (Amsterdam, Rotterdam, and The Hague), represented 46 HIV-infected heterosexual persons of Surinamese (n = 25) and Antillean origin (n = 21) living in The Netherlands, and 9 indigenous Dutch heterosexuals who had no apparent link to Suriname or the Antilles. To study heterosexual HIV transmission among these ethnic groups in The Netherlands, we also obtained 145 pol sequences from the ATHENA national observational cohort (HIV Monitoring Foundation, HMF), of which 62 were from heterosexual HIV-infected persons of Surinamese or Antillean ethnic origin living in The Netherlands.¹³ The remaining 83 pol sequences represented heterosexually infected patients of Dutch ethnic origin who had no apparent epidemiologic link to Suriname or the Antilles. For all samples and sequences, epidemiological data were collected including sex, ethnicity, country of birth, calendar year of HIV diagnosis, presumed country of infection, and age at diagnosis. In compliance with regulatory requirements, written informed consent for participation in the ATHENA study cohort was obtained from all research subjects.

Sequencing and phylogenetic analysis

The procedures of viral isolation, reverse transcription and amplification, and direct (double-stranded) sequencing were described earlier in detail.¹⁴ In our study, three genomic regions were amplified and sequenced: env gp120 C2–V3 region, gag (p17 and partial p24), and pol [protease and half reverse transcriptase (RT)]. Molecular Evolutionary Genetics Analysis version 3.1 computer software was used to generate phylogenetic trees, using the neighbor-joining method.¹⁵ The following distance matrices were generated by the Tamura–Nei substitution model with gamma distribution: pol α = 0.25; gag α = 0.25; env-V3 α = 0.38.^16,17 Bootstrap analysis was conducted on 1000 replicas to test the stability of the clustering. Bootstrap values >70% were considered definitive for significant clustering.¹⁸

A pol tree was constructed of all pol sequences, including those of heterosexually HIV-infected patients of Dutch ethnic origin. Reference strains from subtypes A to J were included in the pol tree, and sequences that showed high bootstrap values to the references were considered as the subtypes circulating, due to the close clustering in that tree. Reference strains were derived from the GenBank database, accession numbers A19UG037, B-US86JRFL, C-95IN21068, DJ258, F193BR020_1, G-FI-HH8793, H-BE-V1997, and J-94-SE7022. New sequences as obtained in our study will be provided on request. Additionally, to determine how HIV strains from Suriname and the Antilles might be linked to HIV strains found in Surinamese and Antillean migrants living in The Netherlands (i.e., clusters of genetically related sequences), phylogenetic trees of env and gag sequences were constructed for persons living in Suriname or the Antilles and for persons of Surinamese and Antillean ethnic origin living in The Netherlands.

Statistical analysis

In all, 127 serum samples were collected in Suriname, the Antilles, and The Netherlands (excluding the 145 pol sequences of the HMF). Sociodemographic and laboratory characteristics for individuals having or not having a positive polymerase chain reaction (PCR) were compared using the t-test, Mann–Whitney U test, Chi-square test, and Fisher's exact test. Data were analyzed using SPSS 14.0, and p < 0.05 was considered statistically significant.

HIV migration events

To estimate the mobility of HIV-1 among persons living in Suriname and the Antilles, and Surinamese, Antillean, and indigenous Dutch persons living in The Netherlands, all subtype B partial RT sequences were used (n = 220). These sequences fall into five groups: (1) indigenous Dutch living in The Netherlands (n = 88), (2) Surinamese living in Suriname (n = 19), (3) Surinamese living in The Netherlands (n = 41), (4) Antilleans living in the Antilles (n = 26), and (5) Antilleans living in The Netherlands (n = 46).

Phylogenetic analysis was accomplished using the neighbor-joining method; branch lengths were estimated by maximum likelihood (ML) under the HKY model using a γ distributed rates heterogeneity among sites as implemented in PAUP*.¹⁹ The tree was converted to midpoint-rooted and applied to estimate the migration events among the selected HIV-1 sequences, by using the cladistic approach described earlier by Slatkin and Maddison.¹¹ The nodes of the trees were assigned with a character according to the ethnic group and the country of residence (e.g., 1–5 for the groups above). The algorithm reconstructs “ancestral” states at each internal node by the criterion of parsimony and estimates the minimum number of changes of state needed to give rise to the observed pattern of states. Given that states have been assigned to the “origin” of the HIV sequences under study, the state changes correspond to HIV migration events. The parsimony analysis of the state changes across the phylogenetic trees was performed by using the MacClade program.²⁰ If this number of state changes between the groups of sequences remains low, the possibility for migration events between these particular groups remains low.

To assess if the observed migration events were statistically different (p < 0.05) from those expected under the null hypothesis of geographic panmixis (i.e., equal probability of mixing), we estimated the number of expected migration events after randomization of the states at the nodes of the tree. The process was repeated in 50 replicas, and the distribution of migration events for all trees gave the expected number of changes.

Results

Patients' characteristics

Table 1 shows the characteristics from the total study population for whom HIV-1 sequences were obtained (n = 272), by country of residence. A larger proportion of blood samples collected in Suriname and the Antilles was from men compared to those collected in The Netherlands. Patients living in the Antilles and The Netherlands were older at diagnosis (median age 47 and 41 years, respectively) compared to patients living in Suriname (median age 35 years). With the exception of two patients originating from Suriname who reported potential risk through blood transfusion, the majority of the patients reported unprotected heterosexual contact as the major risk for HIV infection. One-third (39/108) of the Surinamese and Antillean patients living in The Netherlands believed they were infected with HIV in their country of origin; 35% (38/108) of these patients believed they were infected in The Netherlands, and for 29% (29/108) the presumed country of infection was unknown. Whereas all patients living in Suriname and the Antilles were diagnosed with HIV after 2000, almost half of the Surinamese, Antillean, and Dutch patients living in The Netherlands were diagnosed before 2000 (Table 1).

Table 1.

Characteristics of the Study Population by Country of Residence ^a

	Country of residence
	Suriname	N. Antilles	Netherlands
	Surinamese (n = 35)	Antillean (n = 37)	Surinamese/Antillean (n = 108)	Dutch (n = 92)	Total (n = 272)
Male gender	24 (69)	25 (68)	57 (53)	42 (46)	148 (54)
Median age at diagnosis (IQR)	35 (33–46)	47 (42–57)	40 (34–47)	43 (36–53)	42 (35–50)
Country of birth
Netherlands	—	—	—	92 (100)	92 (34)
Suriname	35 (100)	—	56 (52)	—	91 (33)
Netherlands Antilles	—	37 (100)	52 (48)	—	89 (33)
Sexual orientation
Heterosexual	25 (71)	35 (94)	108 (100)	92 (100)	260 (96)
Bisexual	1 (3)	1 (3)	—	—	2 (1)
Unknown	9 (26)	1 (3)	—	—	10 (4)
Self-reported country of infection
Netherlands	—	—	38 (35)	87 (95)	125 (46)
Suriname	35 (100)	—	16 (15)	—	51 (19)
Netherlands Antilles	—	33 (89)	23 (21)	—	56 (21)
Other	—	—	1 (1)	4 (4)	5 (2)
Unknown	—	4 (11)	30 (28)	1 (1)	35 (13)
Calendar year HIV diagnosis
Median	2003	2002	2000	1998	2001
IQR	(2003–2004)	(2000–2004)	(1997–2003)	(1993–2003)	(1997–2003)

Data are n (%) of patients, unless listed otherwise. IQR, interquartile range.

From the 127 samples collected in Suriname, the Antilles, and The Netherlands, HIV-1 RNA was successfully amplified for 97 env, 106 gag, and 75 pol fragments. For the env V3 sequence a fragment of 276 basepairs (bp) was amplified and analyzed, while a 750-bp fragment for the gag and a 1210-bp fragment for the pol gene were sequenced. The env fragment contains more sequence variation than either gag or pol. This variation can interfere with the amplification efficiency. The low observed polymerase chain reaction (PCR)-positive pol fragments may be related to the length of the amplified product, although the pol gene is the most conserved of the three analyzed gene fragments. Overall, PCR failure was significantly higher for samples collected in The Netherlands (27%) compared to those collected in Suriname and the Antilles (8% and 3%, respectively) (Table 2).

Table 2.

Polymerase Chain Reaction (PCR) Failure for Serum Samples Collected in Suriname, The Netherlands Antilles, and The Netherlands Between 2003 and 2005 ^a

		PCR failure
	N	pol	gag	env	Three regions
Country of sample collection (ethnicity)
Suriname (Surinamese)	35	16 (46)	4 (11)	1 (3)	1 (3)
Netherlands Antilles (Antillean)	37	11 (30)	1 (3)	12 (32)	—
Netherlands (Dutch)	46	21 (46)	11 (24)	8 (17)	2 (4)
Netherlands (Surinamese/Antillean)	9	4 (44)	5 (56)	9 (100)	1 (11)
Total	127	52 (41)	21 (17)	30 (24)	4 (3)

Data are n (%) of serum samples. Data on PCR failure of pol sequences of 62 Surinamese or Antillean and 83 Dutch HIV-infected heterosexuals from the national observational cohort in the Netherlands are not included. The number of samples for indigenous Dutch patients was small due to the low number of pol sequences of indigenous Dutch patients registered at the national observational cohort.

Epidemiological links between countries and ethnic groups

A phylogenetic tree showing the genetic relatedness of all pol sequences including those of heterosexual patients of Dutch ethnic origin (n = 220) is presented in Fig. 1. In this pol tree, the predominant subtype B (98%) was accompanied by a few other subtypes: subtype G (n = 1), subtype C (n = 1), subtype H (n = 1), and two strains that fell between clusters (potential recombinants). These non-B strains were primarily (n = 4) from Antillean patients living in The Netherlands who reported The Netherlands as the presumed country of infection. No data on the sexual partner were available to verify the origin of the non-B infection.

FIG. 1.

Neighbor-joining tree of 220 pol sequences, including those of Dutch heterosexual patients. Surinamese living in Suriname Surinamese living in The Netherlands Antillean living in The Netherlands Antilles Antillean living in The Netherlands Dutch living in The Netherlands.

This phylogenetic tree of pol sequences shows five clusters of related B-viruses: two Surinamese (clusters 1 and 2) and three Antillean (clusters 3, 4, and 5). Clusters 1 (n = 8) and 2 (n = 5) contain sequences of Surinamese patients living in both The Netherlands and Suriname. Of the Antillean clusters, cluster 3 consists of three sequences derived from persons living in the Antilles, while clusters 4 and 5 contain sequences of patients living in The Netherlands and the Antilles. These clustering patterns were maintained in env and gag trees (not shown), although the bootstrap values were lower for three clusters in the env tree and for one cluster in the gag tree, perhaps due to increased variability. In the virus genome, the env gene displays the highest degree of variation.^17,21 However, compared to the pol tree, the env tree had an additional Antillean cluster and the gag tree had an additional Surinamese cluster (data not shown). The reason that these two clusters could not be distinguished in the pol tree is that sequences from the individuals in these clusters were not available for pol due to PCR failure.

No sequences obtained from indigenous Dutch heterosexual patients clustered significantly with any of the Surinamese or Antillean sequences, but were dispersed among the sequences of patients of Surinamese and Antillean ethnic origin.

HIV migration events

This analysis provided additional information about the mobility of HIV-1 subtype B among the ethnic groups sampled. The migration matrix with the observed migration events involving HIV-1 sequences in The Netherlands, Suriname, and the Antilles is shown in Table 3. A low number of HIV migration events indicates little HIV mobility among groups, while the higher the number, the more mobility or relationship there is. Table 4 shows the differences between the observed and the expected values with regard to each ethnic group. Positive and negative values indicate, respectively, higher and lower HIV migration than expected under the hypothesis of panmixis. These figures provide a measure of the levels of genetic migration or isolation for each group. The observed HIV migration from one group to another was significantly lower than expected, except that Antilleans living in The Netherlands had significant higher migration than all the other groups. Similarly, significant higher migration than expected was detected from Surinamese in The Netherlands to the same ethnic group living in Suriname. Migration from the Surinamese in The Netherlands to patients of Dutch ethnic origin, on the other hand, was not significant.

Table 3.

Number of Observed Migration Events ^a Calculated Between HIV-1 Sequences Sampled in The Netherlands, Suriname, and The Netherlands Antilles Between 2003 and 2005 ^b

			Destination
			Netherlands
	Country of sample collection	Ethnicity	Dutch	Antillean	Surinamese	Suriname	Netherlands Antilles
Origin	Netherlands	Dutch		9.6	10.7	2.3	4.8
		Antillean	5.6		12.1	3.8	12.1
		Surinamese	3.9	1.8		5.1	1.1
	Suriname		0.6	0.3	2.4		0.0
	Netherlands Antilles		0.9	2.6	0.3	0.0

These numbers correspond to the HIV migration events estimated among the different ethnic groups and, therefore, correspond to the particular size for each population.

Read migration events out from “origin” (row) to “destination” (column).

Table 4.

Differences between the Observed Migration Events and the Expected Number of Changes Under the Null Hypothesis of Geographic Panmixis for HIV-1 Sequences Sampled in The Netherlands, Suriname, and The Netherlands Antilles Between 2003 and 2005 ^a

			Destination
			Netherlands
	Country of sample collection	Ethnicity	Dutch	Antillean	Surinamese	Suriname	Netherlands Antilles
Origin	Netherlands	Dutch		−16.9	−12.3	−9.1	−11.2
		Antillean	1.0		7.1	1.4	9.0
		Surinamese	−0.1	−2.6		2.9	−1.7
	Suriname		−0.7	−1.1	−1.0		−0.9
	Netherlands Antilles		−1.3	−1.8	−1.8	−1.1

Italics, significantly (p < 0.05) higher migration than expected under the null hypothesis. Bold, significantly (p < 0.05) lower migration than expected under the null hypothesis.

Notably, no HIV migration (number of migration events = 0) was detected between the Surinamese and Antillean ethnic groups living in their home countries (Table 3). The overall number of observed migration events (80) was significantly (p < 0.05) smaller than the expected number of migration events (average value: 115.6) under the null hypothesis of panmixis. This result suggests that there is considerable clustering of the HIV-1 subtype B sequences within ethnic group.

Discussion

Our main objective was to evaluate the pattern of HIV transmission between Suriname, the Antilles, and the Surinamese and Antilleans living in The Netherlands. Phylogenetic analysis showed that subtype B is the predominant HIV-1 subtype circulating among heterosexual patients in The Netherlands Antilles and Suriname, as it was detected in almost all samples in our study. Patients infected with a non-B HIV-1 strain reporting The Netherlands as the presumed country of infection might have acquired these strains by sexual contact with persons originating from areas of high HIV-1 prevalence, where non-B genetic forms predominate,e.g., sub-Saharan Africa or Southeast Asia.^8,22

In the phylogenetic trees of pol, gag, and env sequences, various clusters were observed only for Surinamese or Antillean patients, either those living in The Netherlands or their country of origin. These clusters suggest various Surinamese and Antillean “founder” viruses that have been introduced into The Netherlands from Suriname and the Antilles or from The Netherlands into those areas. The clustering in the phylogenetic analysis supports the results of epidemiological research among Surinamese and Antillean migrants who are traveling regularly to their country of origin. Of the Surinamese and Antillean travelers we studied, 10% reported unprotected sex in both their country of origin and The Netherlands, making them a potential bridge population for transmission of HIV and STIs between The Netherlands and the former colonies.²³ On the other hand, persons living in Suriname and the Antilles travel frequently to The Netherlands and indigenous Dutch persons travel frequently to Suriname and the Antilles. However, no information on sexual contacts while traveling is available for these groups of travelers.

Although epidemiological studies show that heterosexual relationships are not uncommon between Surinamese, Antillean, and Dutch individuals living in The Netherlands,^9,10,24 our phylogenetic tree of pol sequences shows almost no transmission of HIV infection from one group to another. This is in line with epidemiological studies that show sexual ethnic mixing, but also show that condom use is more likely to be consistent in disassortive relationships (i.e., partners of different ethnic origin) than in assortive partnerships (i.e., partners of the same ethnic origin).^9,10 The limited heterosexual transmission of HIV between these ethnic groups might also be explained by the relatively low HIV prevalence among Surinamese, Antillean, and indigenous Dutch in The Netherlands. Whereas the above analysis provides insight into the phylogenetic relationships between HIV-infected individuals of different ethnic origins, an additional analysis sheds light on the phylogenetic relationship between HIV-infected populations. By estimating migration events on the population level, we could study the genetic mobility of HIV among countries and ethnic groups. Our results show high HIV mobility between Surinamese living in The Netherlands and Surinamese living in Suriname and also between Antilleans living in The Netherlands and Antilleans living in the Antilles. We observed no mobility, on the other hand, between patients living in Suriname and the Antilles, suggesting high levels of HIV subtype B isolation in those areas.

These results are in accordance with our epidemiological and phylogenetic analysis in different gene regions. However, the migration matrix suggests HIV mobility between the Antillean ethnic group and all the other ethnic groups under study living either in or outside The Netherlands. Higher, but nonsignificant, rates of disassortive mixing among Antilleans compared to other ethnic groups in The Netherlands were also found in a study among ethnic groups in The Netherlands.²⁴ However, HIV mobility between the Antillean ethnic group in The Netherlands and persons in Suriname could not be explained by earlier studies.

The number of estimated migration events corresponds to the number of individuals within each study population. Therefore, the number of migration events for the Dutch individuals studied was higher, as the Dutch population in The Netherlands was larger than the Surinamese and Antillean population sampled in The Netherlands. However, this was significantly lower than expected. In addition, it is possible that the indigenous Dutch patients registered by the HMF include Surinamese or Antillean migrants of the second generation, because for adults the HMF registers only the patient's country of birth. Taking these considerations into account, the phylogenetic analysis at the population level indicates that HIV transmission among these three ethnic groups in The Netherlands does occur on a limited scale.

For those migrants who reported that they probably acquired their infection in the country of origin, no information was available on whether it was acquired before or after moving to The Netherlands. Infection after migration might have been acquired while traveling to the country of origin. We should also note that the sample collection relied mainly on convenience sampling, although we randomly selected samples when multiple cases were available. Such random selection resulted in a sample of indigenous Dutch patients who were earlier diagnosed with HIV than were patients of Surinamese and Antillean ethnic origin. However, diagnoses of HIV infection might occur earlier after infection in The Netherlands, compared to Suriname or the Antilles, due to more extensive promotion of HIV testing. Furthermore, those of Dutch origin who were diagnosed after 2000 (n = 41, 50%) also showed no clustering with patients of Surinamese and Antillean ethnic origin.

Earlier HIV diagnosis among those living in The Netherlands might account for the higher frequency of PCR failure among Dutch patients in The Netherlands. No HIV RNA levels were available, but if levels are generally lower in Dutch patients, they might be another reason for PCR failure.

This molecular epidemiology study showed that HIV transmission in two large migrant groups in The Netherlands occurs predominantly between migrants living in The Netherlands and their corresponding populations in the country of origin. Transmission between these migrant groups and the indigenous Dutch population rarely occurs. Higher HIV migration than expected was observed only among Antilleans living in The Netherlands. These results suggest that cultural factors could have a stronger influence on the HIV mobility pattern of the populations studied than the geographic distance between ethnic groups. These patterns might change with increasing international travel, migration, and a longer history of migration. Therefore, further molecular epidemiological studies should be continued among HIV-infected indigenous Dutch patients, the first and second generations of HIV-infected migrants in The Netherlands, and HIV-infected patients in their countries of origin to observe a change in patterns.

Footnotes

Acknowledgments

The study was supported by a grant from the AIDS Foundation, The Netherlands (Grant 7015). Dimitorios Paraskevis was supported by the Hellenic Center of Infectious Diseases Control and Prevention (H.C.D.C.P, http://www.keel.org.gr) of the Ministry of Health & Welfare. We thank Anneke van Ijperen, Astrid van Hes, Tessa Royaards, and Mahesh Algoe for assistance in the collection of the epidemiological data, Fokla Zorgdrager, Remco van den Burg, and Radi Hamidjaja for their technical support, Ronald Geskus for critically reading this manuscript, and Lucy Phillips for editing the final manuscript.

Author Disclosure Statement

No competing financial interests exist.

References

Koedijk

FDH

, Vriend

, van Veen

et al. Sexually Transmitted Infections, Including HIV, in The Netherlands in 2008. National Institute for Public Health and the Environment: Bilthoven, 2009.

Vaughan

, Cane

, Pillay

et al. Characterization of HIV type 1 clades in the Caribbean using pol gene sequences. AIDS Res Hum Retroviruses, 2003; 19:929–932.

UNAIDS: Report on the global HIV/AIDS epidemic 2008. UNAIDS: Geneva, 2008.

Op de Coul

, van Sighem

, van de Laar

. National estimate of people living with HIV/AIDS in 2005 in The Netherlands. Infect Dis Bull, 2006; 17:398–403.

Kramer

, van den Hoek

, Coutinho

, Prins

. Sexual risk behaviour among Surinamese and Antillean migrants travelling to their countries of origin. Sex Transm Infect, 2005; 81:508–510.

Van Veen

, Beuker

, de Brito

et al. HIV-Surveys among High Risk Populations in Rotterdam 2002–2003. National Institute for Public Health and the Environment: Bilthoven, 2005.

Van Veen

, Wagemans

MAJ

, van den Burg

et al. HIV-Surveys among High Risk Populations in The Hague 2005. National Institute for Public Health and the Environment: Bilthoven, 2006.

Op de Coul

, Coutinho

, van der Schoot

et al. Dutch HIV-1 subtype surveillance. The impact of immigration on env HIV-1 subtype distribution among heterosexuals in The Netherlands influx of subtype B and non-B strains. AIDS, 2001; 15:2277–2286.

Gras

, Weide

, Langendam

et al. HIV prevalence, sexual risk behaviour and sexual mixing patterns among migrants in Amsterdam, The Netherlands. AIDS, 1999; 13:1953–1962.

10.

Van Veen

, Wagemans

MAJ

, Op de Coul

ELM

et al. HIV-Surveys among High Risk Populations in Amsterdam 2003–2004. National Institute for Public Health and the Environment: Bilthoven, 2005.

11.

Slatkin

, Maddison

. A cladistic measure of gene flow inferred from the phylogenies of alleles. Genetics, 1989; 123:603–613.

12.

Paraskevis

, Pybus

, Magiorkinis

et al. Tracing the HIV-1 subtype B mobility in Europe: A phylogeographic approach. Retrovirology, 2009; 6:49.

13.

HIV monitoring foundation. http://www.hiv-monitoring.nl.

14.

Op de Coul

, van den Burg

, Åsjö

et al. Genetic evidence of multiple transmissions of HIV type 1 subtype F within Romania from adult blood donors to children. AIDS Res Hum Retroviruses, 2000; 16:327–336.

15.

Molecular Evolutionary Genetics Analysis. http://www.megasoftware.net.

16.

Tamura

, Nei

. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol, 1993; 10:512–526.

17.

Leitner

, Kumar

, Albert

. Tempo and mode of nucleotide substitutions in gag and env gene fragments in human immunodeficiency virus type 1 populations with a known transmission history. J Virol, 1997; 71:4761–4770.

18.

Hillis

, Bull

. An empirical test for bootscanning as a method for assessing viral sequence relatedness. System Biol, 1993; 42:182–192.

19.

Swafford

. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), 4th. Sinauer Associates: Sunderland, MA, 1998.

20.

Maddison

, Maddison

. MacClade Version 4: Analysis of Phylogeny and Character Evolution. Sinauer Associates: Sunderland, MA, 2003.

21.

Hué

, Clewley

, Cane

, Pillay

. HIV-1 pol gene variation is sufficient for reconstruction of transmissions in the era of antiretroviral therapy. AIDS, 2004; 18:719–728.

22.

Aggarwal

, Smith

, Tatt

et al. Evidence for onward transmission of HIV-1 non-B subtype strains in the United Kingdom. JAIDS, 2006; 41:201–209.

23.

Kramer

, van Veen

, Op de Coul

ELM

et al.

Migrants travelling to their country of origin: A bridge population for HIV transmission?

Sex Transm Infect, 2008; 84:554–555.

24.

Van Veen

, Kramer

, Op de Coul

et al.

Disassortative sexual mixing among migrant populations in The Netherlands: A potential for HIV/STI transmission?

AIDS Care, 2009; 21,6:683–691.