Resistance Test Project

Sorry it’s taken so long to update–I’ve just been working a lot with spreadsheets and didn’t want to update until I had something interesting: the results. I’ve been working on this project for the better part of 4 weeks while also helping to develop our system for dosing guides. When Dr. Lowenthal approached us with this project, it took us a while to fully grasp it, so please bear with me as I try to explain it now.

Background: HIV+ patients on HAART (highly active antiretroviral therapy) are given their ARVs (antiretrovirals) in what are called lines, or regimens. All pts start with 1st line (usually AZT, 3TC, NVP) and move to 2nd line when and if they fail first line (DDI and d4T is a common 2nd line combination). 3rd line is also available for patients who fail 2nd line. Drug failure can occur for several reasons, but the primary reason for our patients is poor adherence. Most ARVs are taken every 12 hours to ensure a steady level of the drug in the blood; when a patient does not take his drug every 12 hours, the levels in his blood will drop. This is a major problem given HIV’s ability to generate spontaneous mutations. If a mutation is created that is resistant to the drug that is being improperly taken, major problems can occur. If the patient misses a dose, the mutation can occur, but if the drug is taken again regularly, the virus with the mutation will not be able to reproduce effectively. However, if the doses are on and off, the resistance mutation will be able to build while blood levels of the drug are low and then be able to be resistant to higher levels of the drug.

Drug resistance can be seen in pts when their adherence improves to a sufficient level but the drugs are having no effect (viral load not being suppressed). Once this occurs, when the ARVs are no longer effective and the viral load is rebounding (no longer suppressed), most pts are moved to 2nd line therapy. However, several pts in the clinic were fortunate enough to have a resistance test done, which is very expensive but often very helpful in determining care. The resistance test requires a blood draw, and the virus in the blood is analyzed for certain mutations. Researchers have found which mutations cause resistance to certain drugs, so doctors are able to know which ARVs the patient is resistant to before prescribing 2nd-line drugs. However, it is not feasible for every pt to have a resistance test, given the cost.

Question 1: Our role in this was to look at the records of every patient in the clinic who had gotten a resistance test and to see what the typical mutations were for 1st line regimens so that we could determine the best drugs for 2nd-line therapy. Currently, patients who fail on 1st-line therapy with AZT are automatically given d4T for 2nd-line therapy, because it is assumed that resistance to AZT will be greater. However, Dr. Lowenthal was fairly certain of and wanted to show that both AZT and d4T would be acceptable 2nd-line therapies, as this would give doctors the ability to prescribe either one based on factors of the individual pt, such as side effects.

Question 2: Our second question involved the other drug in 2nd-line regimens, DDI. DDI is infamous for its demanding dosing schedule: it cannot be taken with food, so pts must take it an hour before their other ARVs, many of which must be taken with food. So instead of having to take drugs twice a day at for example 7am and 7pm, these pts must take their drugs at 6am, 7am, 6pm, and 7pm. It is especially difficult to maintain adherence in a child, as a caregiver must be present at all dosing times and remember the schedule. However, there is another ARV, called ABC (abacavir), which works in the same way as DDI but is not the national standard for 2nd-line therapy. Dr. Lowenthal wants that changed. She wanted me to compare the resistance to DDI and ABC in the kids who failed first line therapy to see if ABC can stand up to DDI.

 Methods: Rachel and I went through the files of the 70 or so patients who had had resistance tests while at the Baylor Clinc. We recorded all pertinent information and any possible confounding factors into a “super-mega” spreadsheet (my wording). The most important data came from each pt’s resistance test, which listed all mutations the pt had. For each pt, we inputted the mutations into the Stanford HIV Resistance Database (http://hivdb6.stanford.edu/asi/deployed/hiv_central.pl?program=hivdb&action=showMutationForm), which reported which drugs the pt was resistant to, along with a quantitative scoring of the resistance level to each drug. 

Comment on Data Collection: This was a long process, as we had to wade through lab results (to watch viral load to determine when a pt started failing), but the lab results were rarely in any kind of order. There had been several changes in the record-keeping system at Baylor, so it was difficult to wade through clinic notes, going back and forth from somewhat illegible handwriting to Word document printouts and back. Drug regimens were often hard to trace through the notes, and adherence rates were not well-recorded, especially in the early years. The files were often difficult to find, as the classification system had changed and is now under Meditech. Under this new system, some pts have been assigned 2 numbers, which is a huge problem when Meditech goes down because we are not able to find the 2nd and correct number. We ended up spending several hours looking through the stacks and shelves of pt files by name instead of classification number, which was basically akin to finding a needle in a haystack. On the plus side, we got to know the woman in charge of pt files, Mamase, quite well!

Methods, cont.: After the data collection for this spreadsheet, I began organizing the data in a way that would be conducive to effective data analysis. I first separated the pts by which drug regimen they had been on at the time of failure and put them in separate spreadsheets. I next went through all the printouts of the Stanford results and inputted each pt’s resistance scores to AZT, d4T, DDI, and ABC. To address the first question of AZT vs d4T, I looked at both the absolute resistance scores of and the differences in scores between AZT and d4T for pts who had failed 1st-line therapy on AZT. For the second question, I ran the same analyses between DDI and d4T, but this time using all pts who had failed first line therapy, not just those who failed with AZT.

Results:

Question 1: Pts who failed 1st-line therapy on AZT had similar average resistance scores for AZT and d4T.

 

Avg Resistance Score
AZT 14.6
d4T 15.8

Pts who were more resistant to AZT had an average duration of failure of 14.6 months, while those more resistant to d4T had an average DoF of 9.96 months.

Drug with Higher Resistance

Avg Duration of Failure
AZT 14.6 months
d4T 9.9 months

Question 2: Pts who failed 1st-line therapy on both AZT and/or d4T had slighty higher resistance scores for ABC than for DDI.

 

Avg Resistance Score
DDI 19.8
ABC 25.8

The average difference in resistance scores between ABC and DDI was 7.

Conclusions: The results obtained were similar to the expectations of Dr. Lowenthal. For the AZT vs d4T question, I found that pts who failed 1st-line therapy on AZT had similar resistance scores to the 2 drugs. This graph shows the distribution of resistance scores. The average difference between AZT and d4T scores was only 1.24, which is certainly neglible for such a large scale.  This indicates that AZT may be a viable choice for 2nd-line therapy, which would allow doctors to choose between AZT and d4T based on each patient’s specific needs.

The duration of failure was found to be longer on average for pts who were more resistant to AZT, by about 4.5 months. This shows that AZT may be a feasible option for 2nd-line therapy as long as the pt has not been in virological failure for an extened length of time. This possibility was discussed with Dr. Lowenthal before the analysis, along with the hope that we may be able to find a general cutoff date, but the data did not allow for it.

The results for the second questions were just as Dr. Lowenthal had hoped. Although 54 out of 57 patients had higher resistance scores for ABC and DDI, the average difference was only 7 points.  In fact, 51 out of the 54 patients with higher resistance scores to ABC had a difference of 7 between their DDI and ABC scores.

   Seven points on the resistance score scale is a moderate gap, but given the ease of administration of ABC as compared to DDI, ABC is the better choice for 2nd-line therapy. Pts on DDI often have very poor adherence, which in itself leads to a higher risk for drug resistance. Botswana is considering changing its national guidelines to reflect this, and Dr. Lowenthal was wanting more evidence to support the use of ABC as standard 2nd-line therapy. Although the cohort was rather small (57 patients), the data patterns are very strong.

Future Work: I am currently just cleaning up the data and analyses I have and starting to analyze for confounding factors, such as age and viral load at the time of the resistance test. Dr. Lowenthal has also presented me with several questions related resistance to NRTIs and to TAMS, which are mutations which when clustered together tend to indicate resistance to NRTIs. I have collected the appropriate data and have found the general pattern, but which is not strong, unfortunately. I am also looking at the number of TAMS a pt has vs. his viral load at the time of the resistance test. I have found absolutely no detectable patterns for this data–the points are seemingly just randomly scattered about.

Leave a Reply

You must be logged in to post a comment.