Available via license: CC BY 4.0
Content may be subject to copyright.
1
1Type of article: Research article
2Title: Drivers of HIV-1 transmission: the Portuguese case
3Running title: Drivers of HIV-1 transmission in Portugal
4Andrea-Clemencia Pineda-Peña 1,2*, Marta Pingarilho1*, Guangdi Li3, Bram Vrancken4,
5Pieter Libin4,5, Perpetua Gomes6,7, Ricardo Jorge Camacho4, Kristof Theys4, Ana
6Barroso Abecasis1 †, on behalf of the Portuguese HIV-1 Resistance Study Group
7* Andrea-Clemencia Pineda-Peña and Marta Pingarilho shared first authorship
8
91 Global Health and Tropical Medicine, Instituto de Higiene e Medicina Tropical,
10 Universidade Nova de Lisboa, Lisbon, Portugal
11 2 Molecular Biology and Immunology Department, Fundación Instituto de
12 Inmunología de Colombia (FIDIC) and Basic Sciences Department, Universidad del
13 Rosario, Bogotá, Colombia
14 3 Department of Metabolism and Endocrinology, Metabolic Syndrome Research
15 Center, Key Laboratory of Diabetes Immunology, National Clinical Research Center
16 for Metabolic Diseases, The Second Xiangya Hospital, Central South University,
17 Changsha, Hunan 410011, China
18 4 Clinical and Epidemiological Virology, Rega Institute for Medical Research,
19 Department of Microbiology and Immunology, KU Leuven
20 5 Artificial Intelligence Lab, Department of computer science, Vrije Universiteit Brussel,
21 Brussels, Belgium
22 6 Laboratório de Biologia Molecular (LMCBM, SPC, CHLO-HEM), Lisboa, Portugal
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
2
23 7 Centro de Investigação Interdisciplinar Egas Moniz (CiiEM), Instituto Universitário
24 Egas Moniz, Caparica, Portugal
25
26 †Corresponding Author and reprints:
27 Ana B. Abecasis
28 Address: Rua da Junqueira 100, 1349-008 Lisboa, Portugal
29 Telephone: +351 213 652 600 / ext. 269
30 Email: ana.abecasis@ihmt.unl.pt
31
32 ABSTRACT:
33 Background: Portugal has one of the most severe HIV-1 epidemic in Western Europe.
34 Two subtypes circulate in parallel since the beginning of the epidemic. Comparing
35 their transmission patterns and its association with transmitted drug resistance (TDR)
36 is important to pinpoint transmission hotspots and to develop evidence-based
37 treatment guidelines.
38 Methods: 3599 HIV-1 naive patients collected between 2001 and 2014 were included
39 in the study. Sequences obtained from drug resistance testing were used for subtyping,
40 TDR determination and transmission clusters (TC) analyses.
41 Results: Subtype B transmission was significantly associated with young males, while
42 subtype G was associated with older heterosexuals. Young males infected with
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
3
43 subtype B were more likely to be included in TC. Consistently, a decreasing trend of
44 prevalence and transmission of subtype G in Portuguese originated people was
45 observed. Active TCs were associated with subtype B-infected males residing in Lisbon.
46 TDR was significantly different when comparing subtypes B (10.8% [9.5-12.2]) and G
47 (7.6% [6.4-9.0]) (p=0.001).
48
49 Discussion: TC analyses shows that the subtype B epidemic is active and fueled by
50 young male patients residing in Lisbon and that transmission of subtype G in Portugal
51 is decreasing. Despite similar treatment rates for both subtypes in Portugal, TDR is
52 different between subtypes.
53
54
55 KEYWORDS: HIV, transmission, Transmitted Drug Resistance, Portugal,
56 phylogenetics, epidemiology
57
58
59
60
61
62
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
4
63 1. INTRODUCTION
64 Portugal had one of the highest rates of HIV diagnoses in Europe in 2016, with 10.0
65 diagnoses per 100,000 population [1]. Despite the fact that new diagnoses have
66 decreased within the country in the last years [2], the patterns of HIV-1 transmission
67 remain uncertain. Phylogenetic analyses are powerful tools to understand the
68 dynamics of viral transmission [3–7], to provide insights for designing prevention
69 policies.
70 According to the European and Portuguese guidelines for antiretroviral treatment
71 [8,9], a baseline resistance test should be performed to determine transmitted drug
72 resistance (TDR), which can impact the first-line antiretroviral response [10]. The last
73 nationwide survey was carried out in 2003 and showed 7.8% of TDR [11]. Surveillance
74 of TDR is important for the development of treatment guidelines, especially in
75 Portugal where considerable migration from Portuguese speaking countries occurs,
76 including some African countries where the levels of TDR are increasing along with the
77 recent scaling-up of NRTI and NNRTI based treatments [12,13].
78 The epidemiology of HIV-1 in Portugal is unique in comparison to other European
79 countries. Most of the epidemic is caused by parallel sub-epidemics of subtype B and
80 subtype G [14]. Until 2005, subtype B accounted for approximately 40% of infections
81 and subtype G accounted for 30% [11,14,15]. The present large-scale cohort provides
82 the unique opportunity to compare the temporal evolution of the parallel epidemics
83 of these subtypes in the same country. Herein, we use transmission cluster
84 reconstruction to understand the drivers of HIV-1 transmission in Portugal and its
85 correlation with primary drug resistance: prevalence of TDR and factors associated
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
5
86 with the spread of TDR. The characterization of HIV-1 transmission in the Portuguese
87 epidemic can help to design targeted prevention strategies.
88
89
90 2. PATIENTS AND METHODS:
91 Study Population
92 The protocol was in accordance with the Declaration of Helsinki and approved by
93 Ethical Committees of Centro Hospitalar de Lisboa Ocidental (108/CES-2014). The
94 Portuguese HIV-1 drug resistance database contains clinical and genotype resistance
95 testing data from patients followed up in 22 hospitals located around the country [16].
96 The inclusion criteria for the analysis of TDR was age older than 18 years and no history
97 of antiretroviral treatment between January 2001 and December 2014. This cohort is
98 named PT-naive, hereafter. The genomic data included the protease and the reverse
99 transcriptase (HXB2: 2253-3554) obtained through population sequencing using the
100 ViroSeq assay.
101
102 Drug Resistance Assessment
103 Surveillance drug resistance mutations (SDRM) were defined according to the WHO
104 list [17]. The impact of TDR was evaluated with the HIVdb v.7.0 and Rega v.9.1.0
105 (http://sierra2.stanford.edu/sierra/servlet/JSierra?action=algSequenceInput).
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
6
106
107 Subtyping and Transmission Cluster Analyses
108 HIV-1 subtypes were determined with Rega v3 and COMET v.1.0 [18] [19,20]. Subtype
109 G and CRF14_BG were merged in a single group (named hereafter G dataset), given
110 that: i) this genomic region has the same evolutionary origin for G and CRF14_BG
111 strains; ii) the origin of the CRF14_BG strains occurred in the Iberian Peninsula
112 [14,19,21]; iii) previously, we reported that the two tools and the manual phylogenetic
113 analyses were not conclusive whether the sequences were G or CRF14_BG in the
114 present cohort [17]; iv) there was no recombination breakpoint in the genomic
115 regions of protease and reverse transcriptase in these sequences. A statistical sub-
116 analysis was performed considering only “pure" subtype G strains (excluding
117 CRF14_BG), defined by the concordant assignment of the two subtyping tools [19,20],
118 to evaluate how this affected our findings (Supplementary material).
119 For the TCs analysis, the dataset was complemented with controls retrieved from: (i)
120 the treated population of the Portuguese cohort between 2001 and 2014, (ii) the 50
121 best-matched sequences to each sequence of the total cohort of subtype B and G, as
122 retrieved by BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), (iii) all HIV-1 pol subtype
123 B and G sequences available from Portugal in the Los Alamos database
124 (http://www.hiv.lanl.gov) [22]. Three subtype D or B reference sequences were used
125 as the outgroup. Sequences with low quality, duplicates and clones were deleted. The
126 resulting dataset was aligned with Muscle [23] and verified for codon-correctness
127 using VIRULIGN [24]. To avoid convergent evolution, SDRMs were removed [17]. The
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
7
128 final subtype B and G datasets consisted of 7497 and 4372 sequences, both with a
129 length of 1173 nucleotides (IQR:1173-1173).
130 A Maximum likelihood tree was constructed with the GTR+ 4Γ nucleotide substitution
131 model and 1000 bootstraps, as implemented in RAxML version 7.5.5. 21 The
132 transmission clusters (TC) were identified with Cluster Picker using a threshold that
133 included a genetic distance of 0.045 and ≥ 80% bootstrap replicates [5,25]. To evaluate
134 the effect of the definition of TCs in the results, sensitivity analyses were performed
135 with varying genetic distances (0.015, 0.030, 0.045, 0.060) and bootstrap supports (70,
136 90, 95, 98).
137 TCs identified were confirmed with Bayesian Markov Chain Monte Carlo (MCMC)
138 inference, as implemented in BEAST v1.8.2 [26]. The temporal signal of the TCs
139 datasets was evaluated with TempEst [27]. The uncorrelated log-normal relaxed
140 molecular clock with a discretized GTR substitution model and the Bayesian Skygrid
141 coalescent model were used [28]. Three separate MCMC chains were run for at least
142 100 million generations. Convergence was determined with Tracer using a burn-in of
143 10% (http://beast.bio.ed.ac.uk/Tracer). The maximum clade credibility (MCC) tree
144 was constructed with TreeAnnotator after discarding the burn-in, and visualized with
145 FigTree v1.4.2 (http://tree.bio.ed.ac.uk).
146 The TCs analyses included the following definitions: (i) A pair was defined as exactly
147 two patients included in a TC, one of them from the PT-naive cohort; (ii) a cluster ≥3
148 included three or more patients, with at least one from the PT-naive cohort (iii) a TDR-
149 cluster-≥3 or TDR-pair contain at least one PT-naive patient with a sequence harboring
150 a SDRM; (iv) an onward-TDR-cluster had ≥3 patients with the same SDRM in the
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
8
151 majority of the patients and at least one from the PT-naive cohort with TDR, which
152 suggest onward transmission of TDR; and (v) active-TCs included transmission of HIV-
153 1 or/and TDR that involves at least one PT-naive patient within a time frame of ≤5
154 years. The time frame was calculated as the maximum length of time between the
155 ancestral node and the most recent tip (year 2014) of the MCC trees [29]. As such, a
156 TC could be separated in two or more active sub-clusters, since such sub-clusters may
157 indicate the population which actively transmitted HIV-1 or TDR in the last years.
158
159 Statistical analyses
160 Statistical analyses were performed to understand and compare the dynamics of the
161 subtype B and G sub-epidemics in the PT-naive cohort, specifically the factors
162 associated with transmission of HIV-1B and HIV-1G, independently of TDR; and the
163 factors associated with transmission of TDR. All these analyses were performed within
164 and compared between subtypes. Sensitivity analyses were performed and, if a result
165 was discordant in the sensitivity analysis, the difference is clearly stated throughout
166 the manuscript.
167 The Fisher's exact test or regression techniques were conducted to compare between
168 proportions, while the Mann Whitney U test or the t-test were used to compare
169 between median or mean values for continuous variables, as appropriate. Binomial
170 logistic regression was used to determine the factors associated with each epidemic,
171 TDR, and clustering. The Bonferroni method was used for multiple testing adjustments.
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
9
172 The level of statistical significance was set at 5%. The analyses were performed with
173 the statistical R software v.3.2.1.
174
175 3. RESULTS
176 3.1. Study Population:
177 The PT-naive cohort included a total of 3599 patients, 2042 with subtype B (56.7%)
178 and 1557 (43.3%) with G. The socio-demographic factors are shown in Supplementary
179 Tables 1&2.
180
181 3.2. Subtype B sub-epidemic is associated with young males living
182 in Lisbon
183 Regarding the socio-demographic factors associated with the transmission of the sub-
184 epidemic B versus G in the PT-naive cohort, there were significant differences
185 between the two sub-epidemics for age, gender, risk of transmission, residence in
186 Lisbon, and CD4 count in the univariate analyses (Supplementary Tables-1&2), while
187 in the multivariate analyses younger age (Odds Ratio (OR): 0.83 for every increase of
188 10 years (10-years), 95% Confidence interval: 0.79-0.89, p<0.0001), male (OR: 2.66, 2.28-
189 3.09, p<0.0001) and living in Lisbon (OR: 1.44, 1.25-1.66, p<0.0001) were significantly
190 associated with subtype B infections
191
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
10
192 3.3. Transmission of subtype B is driven by young males
193 There were 497 TCs that included 61.2% of the subtype B PT-naive cohort (Table 1,
194 and Supplementary Tables 1&2). When comparing the cohort outside versus inside
195 TCs of subtype B in the multivariate analysis, individuals inside subtype B TCs were
196 younger (OR10-years: 0.83, 0.76-0.90, p<0.0001) and more frequently male (OR 1.42,
197 1.14-1.78, p=0.001) than individuals outside the TCs, indicating that young males are
198 driving this sub-epidemic in Portugal as we have identified in other European
199 cohorts[5].
200 The number of PT-naive patients included in TCs had a peak in 1999 for clusters ≥3
201 followed by a steady decrease, while the peak in the number of patients in pairs
202 occurred in 2002 followed by an up and down curve (Figure-1A). This peak in 1999
203 includes 38% of the PT-naive cohort, which suggests transmission of HIV was still
204 ongoing for subtype B despite the introduction of HAART.
205
206 3.4. Transmission of subtype G is decreasing in native Portuguese
207 people
208 There were 333 TCs that contained 46.6% of the G PT-naive cohort (Table 1, and
209 Supplementary tables 1&2). None of the socio-demographic or clinical factors were
210 significantly associated with transmission of subtype G. Interestingly, a decreasing
211 trend in the percentage of native Portuguese people included in TCs was observed
212 since 2005 (p=0.006, Figure-2A).
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
11
213 The numbers of PT-naive patients included in TCs had an up and down curve, as
214 clusters ≥ 3 originated more frequently in 2001 followed by a steady decrease, while
215 this peak occurred in 2005 for pairs (Figure-1A). In contrast with subtype B, most PT-
216 naive patients were involved in TCs originated before the introduction of HAART; i.e.
217 G: 45 TCs included 266 patients (17.1%) vs B: 63 TCs with 114 patients (9%) ;(p<0.0001).
218
219 3.5. In the last years, subtype B transmission was predominant and
220 occurred between patients sampled in Portugal
221 In a sub-analysis considering only TCs originating in the last five years of the cohort
222 (active-TCs), 24.1% (120/497) and 14.1% (47/333) TCs were subdivided in smaller
223 active-TCs for subtype B and G, respectively (Table-1). These active-TCs included
224 mainly subtype B patients (75,5%, 286/379 compared to 23.9% (93/379) for subtype
225 G (p<0.0001)). Socio-demographic characteristics of the patients in active-TCs and the
226 total cohort were similar (Supplementary Tables 1&2). Males (OR: 6.56, 3.63-11.85,
227 p<0.0001) and patients living in Lisbon (OR: 2.03, 1.12-3.69, p<0.05) were associated
228 with the active transmission of subtype B when compared with G. Since we completed
229 our cohort with controls retrieved from other databases, it is important to note the
230 active-TCs included mainly controls sampled in Portugal for both subtypes, indicating
231 transmission of HIV-1 predominantly occurs between patients in Portugal.
232
233
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
12
234 3.6. TDR in subtype G occurs more frequently in patients followed-
235 up in hospitals outside the Lisbon area
236 In the subtype B PT-naive cohort, the prevalence of TDR was 10.8% [221/2042; 9.5-
237 12.2%], 4.9% [102/2042; 4.1-6.0] for nucleoside reverse transcriptase inhibitors
238 (NRTIs), 4.7% [96/2042; 3.8-5.7%] for Non-NRTIs (NNRTIs) and 3.9% [80/2042; 3.1-4.8]
239 for protease inhibitors (PIs) (Figure-2B). Regarding the socio-economical and clinical
240 factors, none of the factors were associated with TDR in the multivariate analysis for
241 this subtype.
242 For subtype G, the TDR prevalence in the PT-naive cohort was 7.6% [118/1557, 6.4-
243 9.0], 1.7% [27/1557; 1.2-2.5] for NRTIs, 4.9% [76/1557; 3.9-6.0] for NNRTIs and 2.2%
244 [34/1557; 1.6-3.0] for PIs. Older age, heterosexual transmission and living outside of
245 Lisboa and Vale do Tejo regions were significantly associated with TDR in subtype G
246 (Figure-2B). The multivariate analysis showed people living outside of the Lisboa and
247 Vale do Tejo regions associated with TDR in subtype G (OR: 1.87, 1.22-2.88).
248 When subtype B and G were compared, there were higher prevalence of TDR (OR:
249 1.47, 1.17-1.89, p=0.001), NRTIs TDR (OR: 2.98, 1.92-4.76, p <0.0001) and PIs TDR (OR:
250 1.83, 1.20-3.83, p=0.003) for subtype B. Subtype B patients with TDR were also older
251 (OR10-years: 0.68, 0.55-0.82, p= 0.001) and more frequently male (OR: 3.10, 1.83-5.24,
252 p<0.0001).
253
254
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
13
255 3.7. Active and onward transmission of TDR for subtype B is driven
256 by males living in Lisbon
257 Eighty-two subtype B TCs had at least one patient from the PT-naive cohort harboring
258 viruses with SDRMs (TDR-TCs, Table 1). The TDR-TCs included 66.1% (n=146/221) of
259 the total number of patients with TDR compared to 43.2% for G (n=51/118; p<0.0001;
260 Supplementary Table-1), indicating more active transmission of TDR in subtype B.
261 When the Portuguese treated population was included as complementary database
262 for the TCs analyses, it was observed that nearly half (n=39/82) of the TDR-TCs for
263 subtype B included at least one treated patient. However, the number of treated
264 patients in subtype B TDR-TCs decreased over time since 2006 (p<0.0001).
265 The origin of subtype B TDR-TCs was mainly between 1999 and 2005: 23 pairs and 26
266 clusters ≥3 represented 60% (n=88/146) of the PT-naive cohort in clusters harbouring
267 viruses with SDRMs (Figure-1B). There were no socio-demographic factors associated
268 with the transmission of SDRMs for subtype B.
269 Twenty-six TDR-TCs originated in the last five years for subtype B (active clusters;
270 Table-1 and supplementary Table-1). This population was similar to the population
271 involved in subtype B active clusters: male (85%, 64/75), living in Lisboa and Vale do
272 Tejo region and <35 years old (both 65.3%). Twenty-four TDR-TCs had evidence of
273 onward transmission of SDRMs, those were mainly thymidine analog mutations
274 (TAMs) and/or NNRTIs SDRMs. Seven out of those 24 TDR-TCs with evidence of
275 onward transmission were still active in the last five years. The characteristics of the
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
14
276 population reflected that male (77%, 20/26) and living Lisboa and Vale do Tejo region
277 (69.2%) still drive the transmission of TDR.
278 As expected, when the two subtype B and G sub-epidemics transmitting drug
279 resistance were compared, subtype B was significantly associated with TDR
280 transmission (OR: 1.75, 1.24-2.49, p=0.0007). Younger age (OR10-years: 0.53, 0.39-0.73,
281 p<0.0001) and males (OR: 4.53, 1.92-10.66 p=0.0005) were consistently associated
282 with transmission of subtype B TDR in all analyses.
283
284 3.8. The onward and active transmission of TDR for subtype G is
285 limited
286 Thirty-one subtype G TDR-TCs included 43.2% (51/118) patients from the PT-naive
287 cohort harbouring viruses with SDRMs (Table 1 and Supplementary Table-2). Nearly
288 half (n=15/31) of the TDR-TCs for subtype G included at least one treated patient.
289 When considering the time origin of TDR-TCs, 35.3% patients of the PT-naive cohort
290 were involved in TDR-TCs originated between 1996 and 1999, followed by 27.5%
291 between 2000-2003 (Figure-1B). Then, mainly pairs including 27.5% and 25.5% of the
292 TDR-patients were originated in 2000-2003 and 2004-2008, respectively. There were
293 no socio-demographic factors or time trends associated with transmission of TDR.
294 When including controls and comparing people transmitting TDR versus without TDR
295 in TCs of subtype G, age (median: 44, IQR: 34-54 versus 37, 31-46, p=0.002), and viral
296 load were significant in the univariate analysis (median: 4.9 Log-copies/mm3, IQR: 4.4-
297 5.7 versus 4.6, 3.9-5.2, p=0.03). However, those were no longer significant in the
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
15
298 multivariate analyses. When TDR-TCs ≥3 were analysed including controls, residence
299 outside of the Lisboa and Vale do Tejo region was significantly associated with TDR
300 within subtype G in the multivariate analysis (OR: 2.96, 1.29-6.79, p=0.01).
301 Interestingly, this geographical pattern was no longer observed in the seven onward-
302 TDR clusters, and from those only one was an active-TC. Unlike subtype B, the socio-
303 demographic factors did not show any clear pattern.
304
305 4. DISCUSSION
306 To our knowledge this is the first study that uses phylodynamics to describe the drivers
307 of HIV-1 transmission in Portugal. Transmission cluster reconstruction has been
308 previously used to understand HIV-1 and resistance transmission patterns in other
309 settings [3–7]. Herein, we combine and compare the information provided by classical
310 statistical analyses of the most complete Portuguese cohort available, stratified by
311 subtypes, with the one retrieved from transmission clusters analyses.
312 Through our detailed analyses of the Portuguese HIV-1 epidemic, we find strong
313 indications that: 1) transmission of subtype B is associated with younger males; 2)
314 transmission of subtype G is decreasing in Portugal and in the native Portuguese
315 population; 3) transmission of drug resistance has different patterns: males living in
316 Lisboa and Vale do Tejo regions drive the active and onward transmission of TDR, while
317 this transmission is limited for subtype G and does not correlate with any socio-
318 demographic factors.
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
16
319 Importantly, active transmission of subtype B in the last years has been driven by
320 males residing in Lisbon. Although a source of uncertainty is the lack of risk factor
321 information for a large part of our cohort, the consistency of our findings in different
322 analyses suggest an important role of MSM living in Lisbon for this sub-epidemic.
323 These results are consistent with other studies in Europe, Brazil or USA, where young
324 MSMs have been identified as the main drivers of subtype B and TDR transmission [3–
325 5,7]. More studies are needed to evaluate how tourism or migration may influence
326 these results.
327 We observed a decline both in the prevalence and in the number of patients present
328 in subtype G TCs in native Portuguese people since 2005 (Figure-2A). This indicates
329 that transmission of subtype G strains is decreasing and that transmission has been
330 limited in the last years. This sub-epidemic was unique in Europe and mainly
331 circulating within Portugal. It was imported from West Africa, it was associated with
332 intravenous drug users (IDUs) [30] and afterwards became also prevalent in
333 heterosexuals. With the introduction of the needle and syringe program in 1993, new
334 infections in IDUs declined and this could be potentially associated with the decrease
335 of viral transmission together with the introduction of HAART in 1996 [2,31]. This
336 finding corroborates how a long-term effective prevention program impacted HIV-1
337 transmission.
338 While we described that TDR levels differ between subtypes, with higher levels for B,
339 the overall TDR remains stable across time, which agrees with the European study
340 SPREAD [32]. TC analyses indicates that onward transmission of TDR is limited and
341 mainly associated with subtype B and with a decreasing proportion of involvement of
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
17
342 treated population since 2006. A higher TDR level for subtype B for NRTIs and PIs could
343 result from several factors: i) The earlier beginning of the treatment for subtype B
344 patients than for subtype G, and/or ii) a lower fitness of G strains in presence of SDRMs,
345 which would cause faster reversion and consequent lower transmissibility level of
346 SDRMs; and/or ii) behavioral patterns affecting the G sub-epidemic dynamics, with
347 slower transmission rates and therefore higher likelihood that SDRM revert before
348 their onward transmission; and/or iv) different treatment strategies for each subtype,
349 which is unlikely because these patients are treated in the same country with similar
350 regimens.
351 Our results should be interpreted with caution due to the lack of information about
352 the time of infection, risk of transmission, country of origin and limited
353 representativeness for the North region of Portugal [2]. The BEST-HOPE project is
354 prospectively collecting recent socio-demographic and behavioral data to complete
355 the picture of the current patterns of transmission in the country [33]. Finally,
356 phylogenetic analyses have intrinsic limitations since it does not provide information
357 about sexual networks and depends on the sampling density[34].
358 In conclusion, we have shown different patterns of transmission of HIV-1 and
359 resistance for the two most important sub-epidemics in Portugal: subtype B and G.
360 Our findings suggest that long-term prevention policies have impacted the
361 transmission of subtype G in Portugal and resulted in decrease of prevalence of this
362 subtype, while subtype B is reflecting the current patterns of HIV-1 transmission that
363 is happening in other European countries.
364
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
18
365
366 AUTHORS’ CONTRIBUTIONS
367 Conceptualization, Andrea Pineda-Peña and Ana Abecasis; Data curation, Andrea
368 Pineda-Peña and Marta Pingarilho; Formal analysis, Andrea Pineda-Peña, Marta
369 Pingarilho, Guangdi Li, Bram Vrancken, Pieter Libin, Kristof Theys and Ana Abecasis;
370 Funding acquisition, Ana Abecasis; Investigation, Andrea Pineda-Peña, Perpetua
371 Gomes, Ricardo Camacho and on behalf of the Portuguese HIV-1 Resistance Study
372 Group; Methodology, Andrea Pineda-Peña, Guangdi Li, Bram Vrancken and Ana
373 Abecasis; Project administration, Ana Abecasis; Resources, Perpetua Gomes,
374 Ricardo Camacho and on behalf of the Portuguese HIV-1 Resistance Study Group;
375 Supervision, Ana Abecasis; Validation, Marta Pingarilho; Writing – original draft,
376 Andrea Pineda-Peña and Ana Abecasis; Writing – review & editing, Andrea Pineda-
377 Peña, Marta Pingarilho, Pieter Libin, Kristof Theys and Ana Abecasis.
378
379 FUNDING
380 This study was supported by European Funds through grant ‘Bio-Molecular and
381 Epidemiological Surveillance of HIV Transmitted Drug Resistance, Hepatitis Co-
382 Infections and Ongoing Transmission Patterns in Europe - BEST HOPE - (project
383 funded through HIVERA: Harmonizing Integrating Vitalizing European Research on
384 HIV/Aids, grant 249697)’; by L’'Oréal Portugal Medals of Honor for Women in Science
385 2012 (financed through L’'Oréal Portugal, Comissão Nacional da Unesco and
386 Fundação para a Ciência e Tecnologia (FCT - http://www.fct.pt)); by FCT for funds to
387 GHTM-UID/Multi/04413/2013; by the MigrantHIV project (financed by FCT:
388 PTDC/DTP-EPI/7066/2014; by Gilead Génese HIVLatePresenters; by the National
389 Nature Science Foundation of China (31571368); by the Innovation-driven Project of
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
19
390 Central South University (2016CX031); by the Fonds voor Wetenschappelijk
391 Onderzoek – Flanders (FWO) grant G.0692.14, and G.0611.09N; by the
392 VIROGENESIS project that receives funding from the European Union’s Horizon 2020
393 research and innovation programme under grant agreement No 634650. The
394 computational resources and services used in this work were provided by the Hercules
395 Foundation and the Flemish Government – department EWI-FWO Krediet aan
396 Navorsers (Theys, KAN2012 1.5.249.12.). K.T. is supported by a postdoctoral grant
397 from FWO.
398
399 ACKNOWLEDGEMENTS
400 We would like to thank the patients and all the members of the Portuguese HIV-1
401 Resistance Study Group:
402 Fátima Gonçalves, Isabel Diogo, Joaquim Cabanas, Ana Patrícia Carvalho, Sandra
403 Fernandes, Inês Costa, Kamal Mansinho, Ana Cláudia Miranda, Isabel Aldir, Fernando
404 Ventura, Jaime Nina, Fernando Borges, Emília Valadas, Manuela Doroana, Francisco
405 Antunes, Maria João Aleixo, Maria João Águas, Júlio Botas, Teresa Branco, José Vera,
406 Inês Vaz Pinto, José Poças, Joana Sá, Luís Duque, António Diniz, Ana Mineiro, Flora
407 Gomes, Carlos Santos, Domitília Faria, Paula Fonseca, Paula Proença, Luís Tavares,
408 Cristina Guerreiro, Jorge Narciso, Telo Faria, Eugénio Teófilo, Sofia Pinheiro, Isabel
409 Germano, Umbelina Caixas, Nancy Faria, Ana Paula Reis, Margarida Bentes Jesus,
410 Graça Amaro, Fausto Roxo, Ricardo Abreu and Isabel Neves.
411
412 Conflicts of Interest:
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
20
413 The authors declare no conflict of interest.
414
415 AVAILABILITY OF DATA
416 Access to the sequences linked to all clinical and demographic data is available
417 through Euresist (http://www.euresist.org), a project aiming to collect and make
418 available data to study HIV drug resistance and viral diversity in Europe.
419
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
21
420 REFERENCES
421 1. European Centre for Disease Prevention and Control; World Health
422 Organization. Regional Office for Europe HIV/AIDS surveillance in Europe
423 2018. 2017 Data.
424 2. Cortes Martins, H.; Aldir, I. Infeção VIH e SIDA: a situação em Portugal a 31
425 de dezembro de 2017; Instituto Nacional de Saúde Doutor Ricardo Jorge, IP:
426 Lisboa, 2018;
427 3. Ragonnet-Cronin, M.L.; Shilaih, M.; Günthard, H.F.; Hodcroft, E.B.; Böni, J.;
428 Fearnhill, E.; Dunn, D.; Yerly, S.; Klimkait, T.; Aubert, V.; et al. A Direct
429 Comparison of Two Densely Sampled HIV Epidemics: The UK and
430 Switzerland. Sci Rep 2016, 6, 32251.
431 4. Esbjörnsson, J.; Mild, M.; Audelin, A.; Fonager, J.; Skar, H.; Bruun Jørgensen,
432 L.; Liitsola, K.; Björkman, P.; Bratt, G.; Gisslén, M.; et al. HIV-1 transmission
433 between MSM and heterosexuals, and increasing proportions of circulating
434 recombinant forms in the Nordic Countries. Virus Evol 2016, 2, vew010.
435 5. Pineda-Peña, A.-C.; Theys, K.; Stylianou, D.C.; Demetriades, I.; SPREAD/ESAR
436 Program; Abecasis, A.B.; Kostrikis, L.G. HIV-1 Infection in Cyprus, the
437 Eastern Mediterranean European Frontier: A Densely Sampled
438 Transmission Dynamics Analysis from 1986 to 2012. Sci Rep 2018, 8, 1702.
439 6. Pineda-Peña, A.-C.; Schrooten, Y.; Vinken, L.; Ferreira, F.; Li, G.; Trovão, N.S.;
440 Khouri, R.; Derdelinckx, I.; De Munter, P.; Kücherer, C.; et al. Trends and
441 predictors of transmitted drug resistance (TDR) and clusters with TDR in a
442 local Belgian HIV-1 epidemic. PLoS ONE 2014, 9, e101738.
443 7. Wertheim, J.O.; Kosakovsky Pond, S.L.; Forgione, L.A.; Mehta, S.R.; Murrell,
444 B.; Shah, S.; Smith, D.M.; Scheffler, K.; Torian, L.V. Social and Genetic
445 Networks of HIV-1 Transmission in New York City. PLoS Pathog. 2017, 13,
446 e1006000.
447 8. Conselho Científico do Programa Nacional para a Infeção VIH/SIDA
448 Recomendações Portuguesas para o tratamento da infeção por VIH-1 e VIH-
449 2 (2016 versão 1.0).
450 9. EACS European AIDS Clinical Society European Guidelines for treatment of
451 HIV-positive adults in Europe version 9.1. 2018.
452 10. Wittkop, L.; Günthard, H.F.; de Wolf, F.; Dunn, D.; Cozzi-Lepri, A.; de Luca, A.;
453 Kücherer, C.; Obel, N.; von Wyl, V.; Masquelier, B.; et al. Effect of transmitted
454 drug resistance on virological and immunological response to initial
455 combination antiretroviral therapy for HIV (EuroCoord-CHAIN joint
456 project): a European multicohort study. Lancet Infect Dis 2011, 11, 363–
457 371.
458 11. Palma, A.C.; Araújo, F.; Duque, V.; Borges, F.; Paixão, M.T.; Camacho, R.;
459 Portuguese SPREAD Network Molecular epidemiology and prevalence of
460 drug resistance-associated mutations in newly diagnosed HIV-1 patients in
461 Portugal. Infect. Genet. Evol. 2007, 7, 391–398.
462 12. UNAIDS. Joint United Nations Programme on HIV/AIDS Global AIDS Update
463 2018 - Miles to Go: Closing Gaps, Breaking Barriers, Righting Injustices.; 2018;
464 13. Ssemwanga, D.; Lihana, R.W.; Ugoji, C.; Abimiku, A.; Nkengasong, J.; Dakum,
465 P.; Ndembi, N. Update on HIV-1 acquired and transmitted drug resistance in
466 Africa. AIDS Rev 2015, 17, 3–20.
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
22
467 14. Abecasis, A.B.; Wensing, A.M.J.; Paraskevis, D.; Vercauteren, J.; Theys, K.; Van
468 de Vijver, D.A.M.C.; Albert, J.; Asjö, B.; Balotta, C.; Beshkov, D.; et al. HIV-1
469 subtype distribution and its demographic determinants in newly diagnosed
470 patients in Europe suggest highly compartmentalized epidemics.
471 Retrovirology 2013, 10, 7.
472 15. Esteves, A.; Parreira, R.; Venenno, T.; Franco, M.; Piedade, J.; Germano De
473 Sousa, J.; Canas-Ferreira, W.F. Molecular epidemiology of HIV type 1
474 infection in Portugal: high prevalence of non-B subtypes. AIDS Res. Hum.
475 Retroviruses 2002, 18, 313–325.
476 16. Libin, P.; Beheydt, G.; Deforche, K.; Imbrechts, S.; Ferreira, F.; Van Laethem,
477 K.; Theys, K.; Carvalho, A.P.; Cavaco-Silva, J.; Lapadula, G.; et al. RegaDB:
478 community-driven data management and analysis for infectious diseases.
479 Bioinformatics 2013, 29, 1477–1480.
480 17. Bennett, D.E.; Camacho, R.J.; Otelea, D.; Kuritzkes, D.R.; Fleury, H.; Kiuchi, M.;
481 Heneine, W.; Kantor, R.; Jordan, M.R.; Schapiro, J.M.; et al. Drug resistance
482 mutations for surveillance of transmitted HIV-1 drug-resistance: 2009
483 update. PLoS ONE 2009, 4, e4724.
484 18. Alcantara, L.C.J.; Cassol, S.; Libin, P.; Deforche, K.; Pybus, O.G.; Van Ranst, M.;
485 Galvão-Castro, B.; Vandamme, A.-M.; de Oliveira, T. A standardized
486 framework for accurate, high-throughput genotyping of recombinant and
487 non-recombinant viral sequences. Nucleic Acids Res. 2009, 37, W634-642.
488 19. Pineda-Peña, A.-C.; Faria, N.R.; Imbrechts, S.; Libin, P.; Abecasis, A.B.;
489 Deforche, K.; Gómez-López, A.; Camacho, R.J.; de Oliveira, T.; Vandamme, A.-
490 M. Automated subtyping of HIV-1 genetic sequences for clinical and
491 surveillance purposes: performance evaluation of the new REGA version 3
492 and seven other tools. Infect. Genet. Evol. 2013, 19, 337–348.
493 20. Struck, D.; Lawyer, G.; Ternes, A.-M.; Schmit, J.-C.; Bercoff, D.P. COMET:
494 adaptive context-based modeling for ultrafast HIV-1 subtype identification.
495 Nucleic Acids Res. 2014, 42, e144.
496 21. Bártolo, I.; Abecasis, A.B.; Borrego, P.; Barroso, H.; McCutchan, F.; Gomes, P.;
497 Camacho, R.; Taveira, N. Origin and epidemiological history of HIV-1
498 CRF14_BG. PLoS ONE 2011, 6, e24130.
499 22. Kuiken, C.; Korber, B.; Shafer, R.W. HIV sequence databases. AIDS Rev 2003,
500 5, 52–61.
501 23. Edgar, R.C. MUSCLE: multiple sequence alignment with high accuracy and
502 high throughput. Nucleic Acids Res. 2004, 32, 1792–1797.
503 24. Libin, P.; Deforche, K.; Abecasis, A.B.; Theys, K. VIRULIGN: fast codon-correct
504 alignment and annotation of viral genomes. Bioinformatics 2018.
505 25. Ragonnet-Cronin, M.; Hodcroft, E.; Hué, S.; Fearnhill, E.; Delpech, V.; Brown,
506 A.J.L.; Lycett, S.; UK HIV Drug Resistance Database Automated analysis of
507 phylogenetic clusters. BMC Bioinformatics 2013, 14, 317.
508 26. Drummond, A.J.; Suchard, M.A.; Xie, D.; Rambaut, A. Bayesian phylogenetics
509 with BEAUti and the BEAST 1.7. Mol. Biol. Evol. 2012, 29, 1969–1973.
510 27. Rambaut, A.; Lam, T.T.; Max Carvalho, L.; Pybus, O.G. Exploring the temporal
511 structure of heterochronous sequences using TempEst (formerly Path-O-
512 Gen). Virus Evol 2016, 2, vew007.
513 28. Gill, M.S.; Lemey, P.; Faria, N.R.; Rambaut, A.; Shapiro, B.; Suchard, M.A.
514 Improving Bayesian population dynamics inference: a coalescent-based
515 model for multiple loci. Mol. Biol. Evol. 2013, 30, 713–724.
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
23
516 29. Yebra, G.; Holguín, A.; Pillay, D.; Hué, S. Phylogenetic and demographic
517 characterization of HIV-1 transmission in Madrid, Spain. Infect. Genet. Evol.
518 2013, 14, 232–239.
519 30. Esteves, A.; Parreira, R.; Piedade, J.; Venenno, T.; Franco, M.; Germano de
520 Sousa, J.; Patrício, L.; Brum, P.; Costa, A.; Canas-Ferreira, W.F. Spreading of
521 HIV-1 subtype G and envB/gagG recombinant strains among injecting drug
522 users in Lisbon, Portugal. AIDS Res. Hum. Retroviruses 2003, 19, 511–517.
523 31. Direção-Geral da Saúde Programa diz não a uma seringa em segunda mão,
524 Kit Prevenção 2013.
525 32. Hofstra, L.M.; Sauvageot, N.; Albert, J.; Alexiev, I.; Garcia, F.; Struck, D.; Van
526 de Vijver, D.A.M.C.; Åsjö, B.; Beshkov, D.; Coughlan, S.; et al. Transmission of
527 HIV Drug Resistance and the Predicted Effect on Current First-line
528 Regimens in Europe. Clin. Infect. Dis. 2016, 62, 655–663.
529 33. Pingarilho, M.; Pineda-Peña, A.-C.; Gomes, P.; Pimentel, V.F.; Libin, P.; Theys,
530 K.; Martins, M.; Vandamme, A.-M.; Camacho, R.J.; Abecasis, A.B.; et al. BEST
531 HOPE - Cohort of newly diagnosed patients in Portugal 2016.
532 34. Brenner, B.; Wainberg, M.A.; Roger, M. Phylogenetic inferences on HIV-1
533 transmission: implications for the design of prevention and treatment
534 interventions. AIDS 2013, 27, 1045–1057.
535
536
537
538
539
540
541
542
543
544
545
546
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
24
547 Figures legends and Table
548
549 Figure 1: Prevalence in percentage of G subtype in Portuguese originated people
550 during the period of the study (years in the y-axis) when considering the ones who
551 were in transmission clusters (A). The light blue shades are the confidence intervals of
552 the proportion of patients for a period of two years (dark blue line). The first period
553 was excluded given the few number of patients but significance did not change. (B)
554 Prevalence (dot) and 95% confidence intervals of the Transmitted drug resistance for
555 subtypes B (red) and G (dark blue) for the PT-naive cohort and for each drug group.
556 Geographical differences were observed for TDR when Lisbon was compared with
557 other regions. Significant differences are shown with an asterisk. Abbreviations: NRTI:
558 Nucleoside reverse transcriptase inhibitors, NNRTI: non-NRTI, PI: Protease inhibitors;
559 TCs: Transmission clusters, vs: versus, %: percentage.
560
561 Figure 2: The most recent common ancestor (MRCA) or the origin of the transmission
562 clusters in years (y-axis) in the PT-naive cohort. The number of naive patients included
563 in the TCs is shown in the y-axis while the year of origin is in the x-axis. The type of
564 clusters is presented as pairs (dashed line) and clusters ≥3 (solid line) for B (dark red)
565 and G-like (blue). (A) The number of patients from the PT-naive cohort is shown
566 according to the MRCA of the TCs (B) The number of patients from the PT-naive cohort
567 is shown according to the MRCA of the TCs with TDR. The Highly Effective
568 Antiretroviral Therapy (HAART) was introduced in 1996. The needle and syringe
569 program (NSP) was introduced in Portugal in 1993. There were several antiretroviral
570 (ARV) drugs introduced in the earlier 2000s such as Tenofovir, Emtricitabine and
571 Protease inhibitors such as Lopinavir/ritonavir, Fosamprenavir/ritonavir and
572 Atazanavir/ritonavir, which increased the regimen options. Since 2007 onwards,
573 potent drugs such as Darunavir (DRV) and the first integrase inhibitors - Raltegravir
574 (RAL) were introduced.
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
25
Table 1: Characteristics of the transmission clusters (TCs) in the general cohort and in the cohort with TDR. Abbreviations: IQR: interquartile
range, n: sample, NS: No significant, TCs: Transmission clusters TDR: transmitted drug resistance, % percentage.
General population
TDR population in TCs
B
G
B
G
Characteristic
n
%
n
%
p-value
n
%
n
%
p-value
Clusters
Number of cohort clusters
497
100
333
100
82
100
31
100
Median size (IRQ)
2
(2-3)
2
(2-3)
NS
3
(2-4)
2
(2-4)
NS
Number of Cluster ≥3
221
44,5
132
39,6
NS
42
51,2
15
48,4
NS
Median size (IRQ)
4
(3-5)
4
(3-5)
NS
4
(3-6)
4
(3-6)
NS
Number of active clusters
120
24,1
47
14,1
0.0003
26
31,7
4
12,9
NS
Median size (IRQ)
2
(2-3)
2
(2-2)
NS
2
(2-3)
2,5
(2-3)
NS
Number of Cluster ≥3
36
7,2
4
1,2
<0.0001
11
13,4
2
6,5
NS
Median size (IRQ)
3
(3-4)
3
(3-3,25)
NS
3
(3-4)
3
(-)
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
26
Number of clusters that suggests onward TDR
Number of Cluster
24
29,3
7
22,6
NS
Median size (IRQ)
-
-
-
-
3
(3-4)
4
(3-5)
NS
Still active
7
8,5
1
3,2
NS
Median size (IRQ)
-
-
-
-
2
(2-2,5)
3
(-)
Patients
Number in cohort clusters
Number of cohort and controls
1636
100
972
100
343
100
100
100
Number of cohort patients
1250
100
726
100
NS
146
100
51
100
NS
Cluster ≥3
Number of cohort and controls
1084
66,3
570
58,6
0.0001
263
76,7
68
68
NS
Number of cohort patients
828
66,2
411
56,6
<0.0001
99
67,8
30
58,8
NS
Number in active clusters
Number of cohort and controls
302
18,5
99
10,2
<0.0001
75
21,9
10
10,0
0.0089
Number of cohort patients
286
22,9
93
12,8
<0.0001
40
27,4
4
7,8
0.0225
Cluster ≥3
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
27
Number of cohort and controls
134
8,2
12
1,2
<0.0001
43
12,5
6
6,0
NS
Number of cohort patients
128
10,2
12
1,7
<0.0001
19
13,0
1
2,0
NS
Number in clusters that suggests onward TDR
Cluster ≥3
Number of cohort and controls
-
-
-
-
-
95
27,7
32
32,0
NS
Number of cohort patients
-
-
-
-
-
72
49,3
21
41,2
NS
Still active
-
-
-
-
-
Number or cohort and controls
-
-
-
-
-
26
7,6
6
6,0
NS
Number of cohort patients
-
-
-
-
-
26
17,8
2
3,9
NS
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
.CC-BY 4.0 International licenseavailable under a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (which wasthis version posted May 30, 2019. ; https://doi.org/10.1101/655514doi: bioRxiv preprint
