121such as overlapping irritable bowel syndrome, infection, bile salt diarrhea, bacterial
overgrowth, or other causes. Once inflammation/active disease has been confirmed,
a trough sample is collected as data on TDM using trough samples is more robust
than at other time points. Trough samples are typically drawn on the day of infusion
or injection, before the dose of the anti-TNF is received. If drug levels are low, with-
out detectable antibodies, one could increase the dose of the medication or reduce
the dosing frequency. If drug levels are within therapeutic range with undetectable
antibodies, one should consider switching to another class of medication with a dif-
ferent mechanism of action, such as vedolizumab. If antibodies are present, the
concentration of the antibodies may influence the decision- making process as fol-
lows: if antibody concentrations are high, most experts would argue that switching
to either another anti-TNF would be appropriate; if the antibody levels are low, there
are some data suggesting that antibodies can be suppressed with the addition of
immunomodulators and/or increase in anti-TNF dosing, with subsequent improved
clinical response and increased drug levels [ 43 ]. While promising that this may be a
useful approach in patients developing anti-drug antibodies, further research is
required to confirm these results in larger cohorts.
Several studies have assessed the clinical utility of reactive TDM, using a similar
algorithm as above. An initial retrospective cohort study by Afif and colleagues
evaluated the clinical impact of this approach in a cohort comprised of 121 patients
with CD, 31 patients with UC, and 3 patients with indeterminate colitis, who were
receiving infliximab [ 44 ]. Seventy-six patients (49%) underwent TDM evaluation
for loss of response, thirty-four (22%) underwent testing for partial response, and
eight (5%) underwent testing for primary non-response. The remaining 37 patients
were tested for several other indications. Thirty-five of the 155 patients were posi-
tive for ATIs. Consistent with the proposed algorithm, 12 patients with ATIs were
transitioned to another anti-TNF, with 11 of 12 noting an improved clinical response.
Dose optimization of the anti-TNF was performed for 6 of 35 patients with ATIs, of
which only 2 had an improved clinical response (p < 0.016). Sixty-three patients
had subtherapeutic infliximab levels; in this subgroup, 29 underwent dose optimiza-
tion, with 86% noting a clinical improvement. Six patients changed to another anti-
TNF, of which 33% had a clinical response (p < 0.016). Collectively, these data
support the proposed approach to the interpretation of drug levels and anti-drug
antibodies. However, this was a retrospective cohort study and the sample size was
small. These results have been supported in several recent additional studies. Vande
Casteele and colleagues employed an HMSA-based assay and were able to distin-
guish between transient antibody formation and persistent antibodies [ 45 ]. In those
with transient antibodies, the rate of clinical response to dose modification was
69%. However, in those with persistent antibodies measured by HMSA, response to
dose modification was only 16% (p = 0.0028). Another retrospective study by
Pariente and colleagues also suggests that there may be a subset of patients with
ATIs who will respond to dose intensification [ 46 ]. In a cohort of 76 patients with
IBD who had lost response, 16 (22.4%) had ATIs. Ten of these 16 patients under-
went dose intensification of IFX, with a 60% response rate, in which there was a
50% response rate among patients with high-titer antibodies.
8 Therapeutic Drug Monitoring of Biologic Agents