Tumor necrosis factor (TNF) α inhibitors have a good safety and efficacy profile in the treatment of psoriasis. However, induction of antinuclear antibodies (ANAs) and onset of autoimmune diseases during treatment have been reported, mainly in rheumatology journals and, to a lesser extent, in dermatology journals.1–10 Studies are generally based on a low number of patients and have short follow-up periods. The significance of these phenomena remains unclear and is open to debate. Adverse drug effects could play a role, although the possibility of a latent underlying disease when the drugs are administered cannot be ruled out.

We decided to perform a study with the primary objective of evaluating the emergence of autoimmune disease, ANAs, extractable nuclear antigen antibodies (ENAs), and double-stranded DNA (dsDNA) antibodies in patients with psoriasis treated with subcutaneous TNFα inhibitors (adalimumab [ADA] and etanercept [ETN]). The secondary objectives of this study were as follows: (1) to verify whether there is an association between the emergence of antibodies or autoimmune disease and the effectiveness of treatment, possible adverse events, and order of administration of the TNFα inhibitor; and (2) to evaluate factors that predict emergence of ANAs and autoimmune diseases.

We performed a retrospective, single-center cohort study of patients seen at the Dermatology Department of Hospital Universitario de la Princesa, Madrid, Spain between February 2004 and March 2016. The study was approved by the local ethics committee (registration no. 27989).

The study population comprised 121 patients. Nine patients were treated with ETN and ADA, 56 were treated with ETN alone, and 56 with ADA alone. Therefore, ADA was used in 65 patients and ETN in 65 (77 men and 53 women). Table 1 summarizes the main baseline characteristics. Patients in the ADA group were taller and had a significantly lower baseline PASI than patients taking ETN. The number of naïve patients was clearly higher in the ETN group, thus explaining the differences in baseline PASI between the groups.

Table 2 and Figure 2 show data from the first year of follow-up, especially with respect to induction of antibodies and development of autoimmune diseases. We can observe that induction of ANAs becomes more frequent as the time on treatment increases: from 12.19% at baseline to 37.5% at 12 months with ETN and from 25% to 50% with ADA. In the case of ADA, the patterns revealed by indirect immunofluorescence varied widely. No autoimmune connective tissue diseases were observed during follow-up.

Figure 2.

Induction of ANAs during the first year of the study in both treatment groups. ANA indicates antinuclear antibody; ADA, adalimumab; ETN, etanercept; NEG, negative; POS, positive.

(0.13MB).

Figures 3 and 4 show how ANAs changed during the evaluation period.

Figure 3.

Changes in ANAs after treatment with adalimumab. The values in the box on the left are from those patients who only had 1 baseline determination and at least 1 other determination in under 1 year, thus enabling them to enter the study but not to be followed up over time. On the right (Changes over time), we can see determination of ANA in patients who were followed up. All patients whose antibodies changed were considered to have undergone determination of ANA at all follow-up points in order to construct the table.

(0.2MB).

Figure 4.

Changes in ANAs after treatment with ETN. See Figure 3 for explanation.

(0.2MB).

The effectiveness analysis revealed differences in favor of patients taking ADA compared with those taking ENT: the absolute mean PASI was lower at 6 and 12 months and overall (Table 3). However, this value is not reliable, since patients taking ADA started with statistically significantly lower baseline PASI values. At no point during the study was an association found between induction of ANAs and a 75% improvement in the PASI (PASI 75) (Table 4). Assessment of the effect of treatment using the Mantel-Haenszel test revealed that treatment did not significantly modify the association between PASI 75 and induction of ANAs at 3, 6, and 12 months or overall.

The only variable significantly associated with the induction of ANA was BMI (P=.024) (Table 5). The risk of induction of ANA fell as BMI increased, with an odds ratio (OR) of 0.867 (95% CI, 0.759-0.991), thus indicating that in the range 20-30 the risk decreases as BMI increases. The association with age did not reach statistical significance (P=.054). Although the overall adjustment was significant (P=.043), the ability of the independent variables to account for the dependent variable was very small (Cox-Snell R,2 0.055; Nagelkerke R,2 0.076). None of the variables studied correlated with the ability to reach PASI 75.

Given the possibility that administration of a TNFα inhibitor could lead to the induction of ANAs during treatment with a second TNFα inhibitor, we analyzed the order of use of ADA (high percentage of second treatments) and ETN (high percentage of patients who were biologic-naïve). However, no significant differences were observed between induction of ANA and previous therapy with biologics (naïve vs second or third TNFα inhibitor) (Table 6). The Mantel-Haenszel common OR was 0.544 (95% CI, 0.229-1.292), thus indicating that there was no reason to reject the independence between the variables analyzed.

The adverse effects analysis showed that during the first year of follow-up in the ADA group, 4 cases were potentially caused by an autoimmune mechanism (1 paradoxical guttate psoriasiform reaction, 2local reactions at the injection site, and 1 case of alopecia areata, Fig. 5).

Figure 5.

Most common adverse events during the first year of treatment. Adalimumab group: Guttate psoriasis (1.53%), infections (6.15%), local reactions (3.07%), A. areata (1.53%). Etanercept group: Guttate psoriasis (3.07%), infections (9.23%, with 1 case of conjunctivitis), local reactions (4.61%). A. indicates alopecia; ADA, adalimumab; ETN, etanercept; psor., psoriasis; reac., reactions.

(0.1MB).

Subsequently, there were 2new cases of guttate psoriasiform reactions at 2 and 3 years and 1 case of urticaria at 5 years. The severe adverse effects recorded included 1 case of cervical and lumbar hernia and 1 fractured humerus, both of which required the patient to be hospitalized. Severe adverse effects also included adenocarcinoma of the head of the pancreas, type 1 papillary renal cell carcinoma, and 1 case of Alzheimer disease.

In the ETN group, the only adverse events during the first year that were considered to be autoimmune in origin comprised 4cases of paradoxical guttate psoriasis and 3 local reactions (Fig. 5). During subsequent follow-up, the only events described were mild respiratory tract infections, 1 abortion, 1 case of Alzheimer-type cognitive impairment, 1 case of chronic obstructive pulmonary disease, 2cataract operations, 2herniorraphies, and 1 case of lumbar-sciatic pain. There were no reports of lupus erythematosus, other connective tissue disease, or vasculitis. The abovementioned cases of potentially autoimmune origin were not associated with induction of ANAs in the study patients.

TNFα inhibitors are of enormous benefit in patients with moderate-to-severe psoriasis whose previous treatment with other options has proven ineffective. However, induction of antibodies and subsequent adverse reactions mean that these agents have to be used with care in patients with autoimmune disease.5,7,9 Viana et al.13 believed that a homogeneous baseline ANA pattern and anti-dsDNA antibodies leave patients at greater risk of developing autoimmune diseases, such as lupus erythematosus.

A review of the literature reveals controversy over the role of previous therapy with a TNFα inhibitor in efficacy and its potential to predispose to induction of ANAs or autoimmune disease during treatment with a second or third TNFα inhibitor.6,14 We found no association whatsoever between these variables and previous therapy with biologics, thus indicating that they could be associated with genetic susceptibility. Given the possible transient increase in antibody titers after streptococcal infection resulting from the structural similarity between streptococcal M protein and keratin 12, some authors recommend verifying whether patients had an infection when their ANA titer was being determined.1

In addition to previous treatment, we analyzed other potential predictors of induction of ANAs, such as association with psoriatic arthritis, sex, and baseline PASI, although no significant associations arose. In the case of age, the association did not reach statistical significance, despite being close (P=.054). We found no references in the literature associating age with induction of ANA in patients treated with biologics. BMI was the only parameter associated with the development of ANA, albeit as a protective variable. However, we have no reasonable explanation for this observation and found no relevant data in the literature.

Recent studies have proposed that induction of antibodies could result from the reduction in concentrations of the cytokines that participate in clearance of nuclear material in apoptosis (as is the case with C-reactive protein), thus leading to a greater predisposition to formation of ANAs and development of autoimmunity. An association has also been reported with abnormalities of cell apoptosis, which could lead to the release of antigens in the nucleosome and formation of anti-dsDNA antibodies in predisposed individuals. Bardazzi et al.1 reported that whereas ADA and infliximab inhibit the transmembrane and soluble fraction of TNFα, ETN inhibits only the soluble fraction, thus potentially leading to less pronounced cell apoptosis and less pronounced induction of antibodies (although this would not explain the fact that induction of ANAs is less prevalent with ADA than with infliximab).1 Nevertheless, these hypotheses have not been confirmed and would depend to a large extent on individual susceptibility to the drug.3,7,10,15

Although induction of ANAs in patients taking biologics seems quite common (33.3%-77% depending on the series),1,4,13 development of autoimmune diseases is rather uncommon in psoriasis (<1%). This finding was reflected in our study: the only situations that potentially had an autoimmune origin during follow-up were 1 case of alopecia areata and various paradoxical psoriasiform reactions. This could be due to the isotypes of the anti-dsDNA antibodies induced by the drugs: while in other autoimmune diseases, it is IgG forms that participate in pathogenesis, therapy with TNFα inhibitors induces IgM and IgA anti-dsDNA.2–4

It is well known that some patients develop resistance to TNFα inhibitors, which is a cause of secondary failures of efficacy. Although the underlying mechanism is unknown, Pink et al.7 proposed that induction of ANAs during treatment could predispose to a failed response to the drug. In contrast, other studies have not shown any association between induction of antibodies and the loss of response to treatment.1,8 Lora et al.15 went so far as to suggest a trend towards a better response to the drug with more rapid development of ANA. In the present study, the Fisher exact test did not demonstrate a correlation between induction of ANAs and clinical response to the drug (PASI 75).

Our study is subject to a series of limitations. First, its design was retrospective. Consequently some data were lost, since not all data were collected at all the visits. Second, the number of patients was limited, there was a fair degree of variability between them, and the duration of follow-up varied. Third, the limitations of the techniques used to detect ANAs and other antibodies could have led to false positives and false negatives.

• •

  Induction of ANAs is significantly increased during treatment of moderate-to-severe psoriasis with ADA and ETN, although it is not accompanied by development of autoimmune disease. In contrast, there is no significant increase in induction of anti-dsDNA or anti-ENA antibodies.

• •

  No correlation was observed between induction of ANAs and effectiveness of treatment, the order in which the drugs are administered, or the appearance of adverse effects. Similarly, it was not possible to demonstrate factors that predicted induction of ANAs, except for BMI (inversely proportional).

• •

  We recommend determination of ANAs and screening for autoimmune diseases before starting therapy with TNFα inhibitors. However, serial, routine determination during follow-up is not necessary, except in those cases where there are signs or symptoms of suspected autoimmune disease.

The authors declare that they have no conflicts of interest