Mutation  FPV/r  IDV/r  NFV  SQV/r  LPV/r  ATV/r  TPV/r  DRV/r 
I93L  0  0  0  0  0  0  0  0 
The frequency of each mutation at position 93 according to subtype and drugclass experience. Data are shown for the 8 most common subtypes. The number of persons in each subtype/treatment category is shown beneath the subtype. Mutations occurring at a frequency >0.5% are shown. Each mutation is also a hyperlink to a separate web page with information on each isolate, including literature references with PubMed abstracts, the GenBank accession number, and complete sequence and treatment records.
Pos  WT  PI Naive Persons   PI Treated Persons 
A 5468  B 35507  C 11611  D 2212  F 1332  G 1895  AE 9323  AG 4724   A 385  B 14950  C 2083  D 522  F 2555  G 340  AE 192  AG 255 
93 
I 
L ^{24} M ^{0.2} V ^{0.1}  L ^{45}  L ^{97}  L ^{7.8} M ^{0.1}  L ^{22} V ^{0.5} M ^{0.2} D ^{0.2}  L ^{4.9} V ^{0.8} M ^{0.2}  L ^{27} V ^{0.9} M ^{0.2}  L ^{4.8} V ^{0.3} M ^{0.2}   L ^{21} M ^{1.3} V ^{0.5} F ^{0.3}  L ^{36} M ^{0.6}  L ^{97} P ^{0.1}  L ^{31}  L ^{36}  M ^{2.1} L ^{2.1}  L ^{12} M ^{1.1} V ^{1.1}  L ^{2.8} M ^{1.6} 

The first row shows the frequency of the mutation in persons who are PInaive (indicated in green). The second row shows the frequency of the mutation in persons who have received one or more PIs. The following rows show the frequency of the mutation in persons who have received only a single PI. Mutation rates that differ significantly between treated and untreated isolates are indicated in yellow.
Mutation  PI  NumSeq  NumMut  % Mutant  p 
I93F  0  72703  16  0.00  
I93F  >=1  21577  9  0.00  0.186 
I93F  APV  72  0   
I93F  IDV  1168  0   
I93F  LPV  1274  0   
I93F  NFV  1186  0   
I93F  SQV  458  0   
I93F  ATV  220  0   
I93F  TPV  0  0   
I93F  DRV  8  0   
Mutation  PI  NumSeq  NumMut  % Mutant  p 
I93L  0  72703  32193  44.20  
I93L  >=1  21577  8923  41.30  0.000 
I93L  APV  72  23  31.90  0.047 
I93L  IDV  1168  399  34.10  0.000 
I93L  LPV  1274  858  67.30  0.000 
I93L  NFV  1186  409  34.40  0.000 
I93L  SQV  458  122  26.60  0.000 
I93L  ATV  220  99  45.00  0.882 
I93L  TPV  0  0   
I93L  DRV  8  4  50.00  0.975 
Mutation  PI  NumSeq  NumMut  % Mutant  p 
I93M  0  72703  95  0.10  
I93M  >=1  21577  129  0.50  0.000 
I93M  APV  72  0   
I93M  IDV  1168  5  0.40  0.019 
I93M  LPV  1274  6  0.40  0.004 
I93M  NFV  1186  9  0.70  0.000 
I93M  SQV  458  1  0.20  0.896 
I93M  ATV  220  0   
I93M  TPV  0  0   
I93M  DRV  8  0   
Mutation  PI  NumSeq  NumMut  % Mutant  p 
I93T  0  72703  11  0.00  
I93T  >=1  21577  3  0.00  0.852 
I93T  APV  72  0   
I93T  IDV  1168  0   
I93T  LPV  1274  0   
I93T  NFV  1186  0   
I93T  SQV  458  0   
I93T  ATV  220  0   
I93T  TPV  0  0   
I93T  DRV  8  0   
Mutation  PI  NumSeq  NumMut  % Mutant  p 
I93V  0  72703  179  0.20  
I93V  >=1  21577  25  0.10  0.000 
I93V  APV  72  1  1.30  0.445 
I93V  IDV  1168  2  0.10  0.828 
I93V  LPV  1274  3  0.20  0.834 
I93V  NFV  1186  0   
I93V  SQV  458  1  0.20  0.724 
I93V  ATV  220  0   
I93V  TPV  0  0   
I93V  DRV  8  0    Mutation pattern data is not available for I93.
A complete summary of additional in vitro susceptibility data for viruses with I93 obtained using other assays including the Antivirogram can be found here.
 Least Square Regression (LSR) was used to learn the relative contribution of each mutation to the fold decrease in susceptibility for an ARV. The figure on the left (click to enlarge the figure) shows the regression coefficients (which correlate with the contribution to resistance) for the 35 nonpolymorphic PIresistance mutations shown to contribute decreased susceptibility to at least one PI. A complete description of the method that generates this figure can be found at Rhee et al PNAS 2006. 
Reference  PreviousPI  FollowUpPI  Other Rx  No.Pts  Weeks  Effect of baseline mutations on response 
Duval(2002)  >=1 PI  APV (1,200 mg BID)vs APV/r (450900 mg / 100 mg BID)  NRTIs + EFV  22  24  Most pts had >= 1 DRM at position 46, 54, 82, or 90. 13/14 receiving APV/r vs 2/8 receiving APV had HIV1 RNA <200 at W24. 
Marcelin(2003)  >=1 PI  APV/r (600 mg / 100 mg BID)  NRTIs  49  12  70% of 49 pts achieved RNA <400. Lack of response was associated with >5 of the following DRM: L10F/I/V, K20M/R, E35D, R41K, I54V, L63P, V82A/F/T/S, and I84V. 
Clevenbergh(2004)  >=1 PI  APV/r(APV 600 mg BID, RTV 100 mg BID)  Highly variable  39  16  No. DRM in the 2002 French ANRS algorithm (L10I, V32I, M46I/L, I47V, I50V, I54L/M/V, G73S, V82A/F/I/T/S, I84V, L90M) correlated with RNA changes. Median RNA decrease was 1.4 logs in 25 pts with <=3 DRM and 0.3 logs in 14 patients with >=4 DRM. 
Lastere(2004)  >=1 PI  APV (1,200 mg BID)  Highly variable  84  12  In univariate and multivariate analyses DRMs at positions 10, 20, 36, 73, 82, and 90, and gag p6 PTAPP insertions were associated with poorer VR. 
Johnson(2005)  >=1 PI  ATV/r (ATV 300 mg; RTV 100 mg QD)  2NRTI including TDF  120  48  At baseline, 15%, 41%, 19%, and 25% of pts had a median of 0, 12, 34, or >4 DRMs at the following positions: 10, 20, 24, 33, 36, 46, 48, 54, 63, 71, 73, 82, 84, and 90. Overall, 38% achieved RNA <50 copies/ml including 44% with <4 DRM and 25% with >=4 DRMs 
Bertoli(2006)  >=1 PI  ATV/r (300 mg / 100 mg) QD; ATV 400 mg QD  OB  74  1224  For ATV/r, L10C/I/V, V32I, E34Q, M46I/L, F53L, I54A/M/V, V82A/F/I/T, and I84V reduced probability of RNA < 50 (92%, 83%, 75%, and 0% when 0, 1, 2, or >=3 mutations were present respectively). For ATV, G16E, K20I/M/R/T/V, V32I, L33F/I/V, F53L/Y, I64L/M/V, A71I/T/V, I85V, and I93L/M reduced probability of RNA < 50 (83%, 67%, 6%, and 0% for 0, 12, 3, or >=4 mutations, respectively). 
Pellegrin(2006)  >=1 PI; median 5 PIs, 6 NRTIs, 1 NNRTI  ATV/r (300 mg / 100 mg) QD  2NRTIs (75%), 1 additional PI (17%); T20 (4%)  71  12  L10F/I/V, K20M/R, L24I, M46I/L, Q58E, L63P, G73S/A, V77I, V82A/F/S/T, I84A/V, L90M were associated with failure to reach RNA <50 copies/ml (p<=0.1) in a univariate analysis. A score that also included I54L/M/T/V, A71I/L/V/T was significantly associated with response in that 63% of persons with <5 total mutations vs 11% with >=5 total mutations had RNA <50 copies/ml. G16E and D60E occurred at baseline in 5 and 9 persons, respectively, but were not associated with VR. 
Vora(2006)  >=1 PI  ATV/r(ATV 300 mg; RTV 100 mg QD)  OB  62  12  13 PI mutations at baseline were associated with a reduced VR: 10F/I/V, 16E, 33F/I/V, 46I/L, 60E, 84V, 85V, and 90M. RNA decrease >1 log occurred in 100% with <2 mutations, 80% with 2 mutations, 43% with 3 mutations, and 0% with 45 mutations. In a followup study of 53 patients (Marcelin 2006), only four mutations (L10F/I/V, L33F/I, I84V, and L90M were predictive of reduced response, although the original score remained predictive. 
De Meyer(2008) POWER  >=1 PI (and >=1 of 30N, 46IL, 48V,50LV, 82AFST, 84V,90M); median 5 PIs, 6 NRTIs, 1 NNRTI  DRV/r (600 mg / 100 mg) BID  OB +/ T20  377  24  V11I, V32I, L33F, I47V, I50V, I54L/M, G73S, L76V, I84V, and L89V at baseline were associated with a decreased VR to DRV/r. About 60% with 0, 45% with 12, and <=20% with >=3 DRMs had RNA <50 copies/ml at wk 24. In phenotypic studies, I50V, I54M, L76V, and I84V reduced susceptibility to the greatest extent. V32I emerged in 30% of failures according to prescribing information. 
De Meyer(2008) POWER followup study combined with DUET1 and 2     467  24  Ten of the 11 mutations (all except G73S: V11I, V32I, L33F, I47V, I50V, I54L/M, G73S, L76V, I84V, and L89V) in the previous De Meyer 2008 study and a new mutation T74P were associated with a decreased VR (defined by RNA <50 copies/ml) at W24. In patients who did not received T20, harboring >=3 of these mutations were associated a decreased VR. 
Pellegrin(2008) PREDIZISTA  >=1 PI; median 5 PIs, 6 NRTIs and NNRTIexperienced (91%)  DRV/r (600 mg / 100 mg) BID  2 NRTIs +/ 1 PI +/ T20  67  12  26(40%) of patients resulted in VF which was defined as RNA >2.3 at W12. I13V, V32I, L33F/I/V, E35D, M36I/L/V, I47V, F53L,I62V at baseline were associated with increased VF. Adjusted OR for resulting in a VF for one addition of these mutations was 6.2. 11% with <4 , 48% with 45 and 100% with >5 these mutations resulted in a VF. 
Marcelin(2007) CONTEXT and TRIAD studies  >=1 PI  FPV/r (700/100 BID)  Usually 2 NRTIs  113  12  The mutations I15V + M46IL + I54LMV + D60E + L63PT + I84V. Persons with 0 or 1 mutation had a mean 2 log decrease, those with 2 mutations had a median 1.5 log decrease, and those with >=3 mutations had <=0.6 log decreases. Mutations at positions 10, 33, 73, and 90 were negatively associated with response in univariate analyses. In this APV/FPVnaive population, no patient had V32I, I47V, or I50V. 
Pellegrin(2007)  Median of 3 (range 1:5); 26% had received APV  FPV/r (700/100 BID)  OB  121  12  10IFRV, 33F, 36I, 46IL, 54LMTV, 62V, 63P, 71ILVT, 73ACST, 82AFST, 84V, and 90M were associated with decreased VR. The presence of <4 mutations was associated with a median 2.3 log decrease whereas the presence of >=4 was associated with a 0.1 log decrease. 
Masquelier(2008)  >=1 PI  FPV/r (700 mg / 100 mg) BID  OB  63  12  L10F/I/V, L33F, M46I/L, I47V, I54L/M/V/A/T/S, A71V, G73C/S/A/T, V82A/F/C/G, and L90M were associated with a decreased VR; V77I and N88S were associated with an increased VR. 
Para(2000)  SQV (>=12M)  IDV (800 mg TID)  NRTIs  51  8  Mutations at positions 10, 20, 48, 82, 84, and 90 predicted a decreased VR at W8. 0 mutations: 1.7 log decrease, 1 mutation: 1.1 log decrease, 25 mutations: 0.3 log decrease. 
Shulman(2002)  IDV ± PIs other than RTV  IDV/r (400 mg of IDV and RTV BID)  NRTIs  31  48  10/14 vs 3/14 subjects with >=3 mutations at the following positions responded to RTV boosting: 10, 20, 30, 32, 33, 36, 46, 47, 48, 50, 54, 71, 73, 77, 82, 84, 88, 90 
Campo(2003)  >= 1 PI(26/28 IDV, 15/28 RTV)  IDV/r (IDV 800 mg and RTV 200 mg BID)  NRTIs ± NNRTIs  28  24  VR correlated better with adherence than the pattern of drugresistance mutations. 
Saah(2003)  NFV  IDV (1000 mg TID)  EFV  29  48  D30N occurred in 17 and L90M in 11. RNA <500 occurred in 9/17 with D30N vs 2/11 with L90M. 
Kempf(2002)  >=1 PI  LPV/r  NRTIs + EFV  50  72  DRMs at 11 positions were associated with decreased VR: 10, 20, 24, 46, 53, 54, 63, 71, 82, 84, and 90. 21/23 with 05, 15/21 with 67 mutations, and 2/6 with 810 of these DRM had RNA <400 at W72. 
Masquelier(2002)  >=2 PIs  LPV/r  Highly variable  68  12  34% of pts had RNA <400. Lack of response was associated with the baseline DRMs M46I, I54V, and V82A and with >=5 of the Kempf DRMs. 
Bongiovanni(2003)  >=1 PI  LPV/r  NRTIs ± NNRTIs  134  12  71/112 pts with <5 and 5/22 pts with >=6 Kempf DRM had RNA <50 at W12. 
Delaugerre(2004)  PInaïve (21); >=1 PI (48)  LPV/r  NRTIs (PInaïve) + NNRTI  69  2472  Among the PIexperienced pts, L10I/F, M46I, I54V/L, A71V/I, V82A/F/S/T, and L90M were associated with VF. 
Loutfy(2004)  >=1 PI  LPV/r  Highly variable  456  412  In univariate analyses, the DRMs most predictive of VF were M46I, Q58E, V82A/F/T, and L90M. The most predictive 3mutation combination was L10F/I/R/V, M46I, and V82A/T/F. 
Marcelin(2005)  Median of 3 (range 15)  LPV/r  OB  116  24  The Genotypic Inhibitory Quotient (GIQ) defined as the median LPV Cmin concentration divided by the number of mutations at the following positions (10, 20, 24, 33, 36, 47, 48, 54, 82, 84). In a multivariate analysis, the GIQ but not the number of mutations was significantly associated with VR. 
King(2007)  3.1 previous PIs (excluding ATV, TPV, DRV)  LPV/r  OB  792  923  Mutations at positions 10, 20, 24, 33, 36, 47, 48, 54, 82, and 84 were associated with decreased likelihood of 1.0 log10 RNA decrease or RNA <400. Mutations at positions 46, 53, 63, 71, 90 from the original score were not significantly associated with response in a multivariate analysis. 
Grant(2008)  LPV/rnaive (100%); ARVexperienced (97%); PIexperienced (79%).  LPV/r  OB  103  24  76% patients achieved VR defined by RNA <500 at W24. Baseline mutations at positions 10, 54 and 82 and at positions 54, 84, 90 were associated with poor VR in univariate and in multivariate analyses, respectively. 
Lawrence(1999)  SQV  NFV (750 mg TID)  NRTI  16  24  No sustained responses were observed. L90M was associated with more rapid VF. 
Casado(2001)  IDV ± RTV  NFV/SQV (NFV 7501250 mg; SQV 6001000 mg BIDTID)  d4T/NVP  31  26 52  35% and 56% of patients had RNA <50 after 6 and 12 months. L90M (0% vs 43%) but not V82A (38% vs 36%) decreased the rate of response. 
Walmsley(2001)  >=1 PI(SQV, IDV, RTV)  NFV (750 mg TID)  NRTIs ± NNRTIs  63  24 48  41% and 22% had RNA declines >=0.5 logs at 24 and 48 weeks respectively. Mutations at position 48, 82, 84, and/or 90 were present in 69% and were correlated with a poor VR. 
Harrigan(1999)  PInaïve (30); >=1 PI (37)  SQV/r (400600 mg of each BID)  NA  67  24  M46I/L, G48V, I54V, A71V/T, V82A/T, I84V, and L90M were weakly associated with a poor VR 
Tebas(1999)  NFV  SQV/r (400 mg of each BID)  NRTIs  24  24  At baseline, 13 had D30N and 5 had L90M. 17/24 had RNA <500 and 10/24 had RNA <50 at W24. Insufficient data to establish association with baseline genotype. 
Zolopa(1999)  >=1 PI  SQV/r (400600 mg SQV; 300400 mg RTV BID)  NRTIs  54  26  RNA decrease of >=0.5 occurred in patients with <= 3 of mutations at positions 46, 48, 54, 82, 84, and 90. D30N did not affect response to SQV/r. 
Marcelin(2004)  >=1 PI  SQV/r (800 mg SQV and 100 mg RTV BID)  NRTIs  72  16  L10F/I/R/V, L24I, M46I/L, G48V, I54V, I62V, A71V/T, V82A/T/F/S, I84V, and L90M were univariate predictors. L24I, I62V, V82A/T/F/S, I84V, and L90M were the best predictors in multivariate analyses: Patients with 0, 1, and >=2 of these 5 mutations had a median 2.2, 1.2, and 0.3 log RNA decrease 
Baxter(2006) RESIST1 (1182,12; NCT00054717) and 2 (1182.48; NCT00144170) + 3 phase II studies (1182.2, 1182.4, 1182,51, 1182.52)  >=2 PIs (and >=1 of 30N, 46IL, 48V,50LV, 82AFLST, 84V, 90M but <3 mutations at codons 33, 82, 84, or 90); median 4 PIs, 6 NRTIs, 1 NNRTI  TPV/r (500 mg / 200 mg) BID  OB +/ T20  688  24  21 mutations at 16 positions were found to correlate with a decreased VR to TPV/r salvage therapy: L10V, I13V, K20M/R/V, L33F, E35G, M36I, K43T, M46L, I47V, I54A/M/V, Q58E, H69K, T74P, V82L/T, N83D, I84V. Each additional mutation was associated with a 0.04 log decreased 2week and 0.16 log decreased 24 week response. The 24 week response dropped from 1.3 logs when 3 mutations were present to 0.64 logs when 4 mutations were present and was completely lost when 8 mutations were present. Note: The vast majority of isolates used to derive the list belonged to subtype B which is relevant because I13V, K20 mutations, M36I, and H69K are highly common in several nonB subtypes. 
Hall(2008) RESIST (phase II/III) followup study  >=2 PIs (and >=1 of 30N, 46IL, 48V,50LV, 82AFLST, 84V, 90M but <3 mutations at codons 33, 82, 84, or 90); median 4 PIs, 6 NRTIs, 1 NNRTI  TPV/r (500 mg / 200 mg) BID  OB +/ T20  688  8/48  In a weighted genotypic susceptibility score (GSS), T74P (weight=7), I47V (6), V82L/T (5), Q58E (5), N83D (4) were the best predictors of poor virologic response; whereas I54A/M/V (3), I84V (2), M36I (2), K43T (2), L10V (1), and M46V (1) were weaker predictors whereas L24I (2), I50L/V (4), I54L (7), and L76V (2) were predictors of virolgic response. 
Marcelin(2008)  Median 4 PIs, 6 NRTIs; NNRTIexperienced (80%); T20experienced (28%)  TPV/r (500 mg / 200 mg) BID  NRTIs +/ T20 +/ NNRTI  688  12  79 (55%) patients achieved VR defined by a decrease of >= 1 log or a BLQ in RNA level at week 12. Baseline mutations at 6 positions found to be associated with a lower VR and one with a higher VR were used for GSS: E35D/G/K/N + M36I/L/V  F53L/W/Y + Q58E + Q61D/E/G/H/N/R + H69I/K/N/Q/R/Y + L89I/M/R/T/V. 100% patients with a GSS of 1, 79% with 0, 56% with 1, 33% with 2, 21% with 3 and 0% with 4 achieved VR. 

 