Is the intake of selective serotonin reuptake inhibitors associated with an increased risk of dental implant failure?


The aim of this retrospective study was to investigate the association between the intake of selective serotonin reuptake inhibitors (SSRIs) and the risk of dental implant failure. Patients were included if they were taking SSRIs only and no other medication, did not present any other systemic condition or compromising habits (bruxism, smoking, snuff), and complied with the use of prophylactic antibiotics for implant surgery. The multivariate generalized estimating equation (GEE) method and multilevel mixed-effects parametric survival analysis were used to test the association between SSRI exposure (predictor variable) and the risk of implant failure (outcome variable), adjusting for several potential confounders (other variables). The total number of implants with information available and meeting the necessary eligibility criteria was 931 (35 failures). These were placed in 300 patients. The implant failure rate was 12.5% for SSRI users and 3.3% for non-users ( P = 0.007). Kaplan–Meier analysis showed a statistically significant difference in the cumulative survival rate ( P < 0.001). The multivariate GEE model did not show a statistically significant association between SSRI intake and implant failure ( P = 0.530), nor did the multilevel model ( P = 0.125). It is suggested that the intake of SSRIs may not be associated with an increased risk of dental implant failure.

Nowadays dental implant placement is an effective and predictable treatment modality for replacing missing teeth in both fully and partially edentulous patients. Nevertheless, failures still happen despite high implant survival and success rates. Several risk factors have been suggested to influence the failure of implants. Surgical conditions, radiotherapy, the oral microbial environment, parafunctional habits, and prosthetic variables are some of these factors. Systemic diseases and compromising risky habits may affect the oral tissues by increasing their susceptibility to other diseases or by interfering with wound healing. The patient’s intake of medications that directly or indirectly affect bone metabolism may also play a role in the outcome of implants.

Among the drugs commonly prescribed today are the selective serotonin reuptake inhibitors (SSRIs). SSRIs are a class of drugs typically used as antidepressants in the treatment of major depressive and anxiety disorders. Studies have shown that the use of antidepressants predicts decreased bone mineral density in women, and both depression and the use of antidepressants are suggested to be possible risk factors for osteoporosis in men. It is possible that neuroendocrine mechanisms related to the serotonin system could regulate osteoclast differentiation/activation, because osteoclasts derive from haematopoietic cell precursors and a relationship between bone and the immune system has been established. Studies have identified a functional serotonin system in osteoblasts and osteoclasts, in which the serotonin transporter and several receptors are expressed in osteoblasts as well as in osteoclasts. The presence of serotonin receptors and the serotonin transporter in bone raises the question whether medications that antagonize serotonin reuptake could influence bone metabolism.

It has been shown in in vitro studies that activity of the serotonin transporter is required for osteoclast differentiation. While blockage of the serotonin transporter was found to reduce osteoclast differentiation when fluoxetine, an antidepressant, was administered to produce micromolar concentrations, there was an increase in osteoclast differentiation for the same medication in the nanomolar concentrations. In vivo studies have demonstrated detrimental effects of fluoxetine on the trabecular architecture and on bone mineral density in mice. Another in vivo study showed that serotonin acts on osteoblasts, inhibiting their proliferation. These animal studies indicate a negative effect of SSRIs on bone mass and suggest that these antidepressants may possess direct anti-anabolic skeletal effects through the pharmacological inhibition of the serotonin transporter.

Therefore, the intake of SSRIs could in theory interfere with the osseointegration process. In the case of dental implants in particular, the findings of recent studies suggest that treatment with antidepressants is associated with an increased risk of failure of osseointegrated implants, while others have not found a relationship between these two factors. Thus, there is still no clear consensus on the influence of antidepressants on the risk of dental implant failure.

As the recognition of conditions that place the patient at a higher risk of failure will allow the surgeon to make informed decisions and refine the treatment plan to optimize the clinical outcome, the purpose of this study was to investigate the association between the intake of SSRIs and the risk of dental implant failure. It was hypothesized that patients taking SSRIs would have a higher implant failure rate than patients not taking this class of drugs. The specific aims of the study were to compare the implant failure rates between users and non-users of SSRIs, and to estimate the influence of several variables on the prevalence of implant failure in regression models, with the intake of these medications as the predictor variable.

Materials and methods

Study design/sample

A retrospective cohort study was designed and implemented to address the research purposes. The study population comprised all patients treated consecutively with implant-supported prostheses between 1980 and 2014 at one specialist clinic (Clinic for Prosthodontics, Centre of Dental Specialist Care, Malmö, Sweden).

To be included in the study sample, the patient had to be taking only SSRIs and no other medication and not present any other systemic condition. The analysis was based on complete cases only; i.e. only those implants with information available for all variables investigated here (see section on Data collection below) were included in the analysis. As it has been suggested that the use of antibiotics in healthy patients significantly decreases early implant failure, all patients had to have taken prophylactic antibiotics for implant surgery in order to be included. All modern endosseous dental implants with a cylindrical or conical design were included.

Patients were excluded as study subjects if they presented a severe systemic disease (American Society of Anesthesiologists physical status III or IV) or had been subjected to irradiation of the head and neck region, were pregnant, alcoholic, bruxers, or smokers, presented a medical disorder known to substantially affect bone metabolism (such as hyperthyroidism, hypothyroidism, vitamin D deficiency, osteomalacia, osteoporosis, Paget’s disease, cancer (excluding non-melanoma skin cancer), diabetes), or were taking corticosteroids, antihypertensive drugs, immunosuppressive drugs, antithrombotic agents (antiplatelet, anticoagulant, thrombolytic drugs), antiepileptic drugs, proton pump inhibitors, bisphosphonates, medications for asthma, or medications to decrease high levels of cholesterol. Thus, the status ‘taking SSRIs’ was isolated as much as possible from the influence of other systemic conditions or medications. Zygomatic implants were not included in the study, nor were implants detected in radiographs but without basic information about them in the patient’s files.

In accordance with the standard protocol at the study clinic, the patients’ dental hygiene was followed up by a dental hygienist within 6 months after the final implant-supported/retained restoration. Each patient then attended a dental hygiene recall programme based on individual needs.

The trial from which data in this study were derived is registered with the US National Institutes of Health ( , NCT02369562).


In this study, the patient’s SSRI status was the predictor variable. SSRI users were defined as patients who reported taking this type of medication during the pre-surgery appointment that was scheduled 1 to 2 weeks prior to implant placement. The SSRIs verified included citalopram, dapoxetine, escitalopram, fluoxetine, fluvoxamine, indalpine, paroxetine, sertraline, venlafaxine, and zimelidine.

The outcome variable was implant failure. An implant was considered a failure in the presence of signs and symptoms that led to implant removal, including lack or loss of osseointegration, implant mobility, continuous pain, advanced marginal bone loss, and refractory infection.

The following factors were the other variables investigated: implant surface (turned/machined or enlarged surfaces, the latter including sandblasted, acid-etched, sandblasted + acid-etched, anodized), implant length (three categories: 6.0–10.0, 10.5–14.0, 15.0–20.0 mm), implant diameter (three categories: 3.00–3.50, 3.70–4.10, 4.20–5.00 mm), prescription of antibiotics (the prophylactic antibiotic regimen was usually started 1–2 h before surgery and continued for 5–7 days postoperatively), bone graft procedures, implant jaw location (maxilla/mandible), anterior or posterior location of the implant (locations 13–23 and 33–43 were considered anterior), patient sex, patient age at implant insertion surgery (three categories: ≤30, >30 to ≤60, >60 years), number of days until failure, and the duration of follow-up.

Data collection methods

The dental records of all patients ever treated with implants at the study clinic were read in order to collect the data. The data were entered directly into an SPSS file as the files were being read (IBM SPSS Statistics version 23; IBM Corp., Armonk, NY, USA). In order to identify patients with the systemic conditions that would exclude them from the analysis, the general health and behavioural history of each patient was collected from the patient files. The presence of a medication list in the patient records was also used to correlate the use of certain drugs to specific health conditions.

Data analyses

With regard to descriptive statistics, the mean ± standard deviation or number and percentage were recorded. Differences between implants placed in SSRI users and SSRI non-users were compared with the Student t -test or Mann–Whitney test for continuous variables, depending on the normality of the data, and with the Pearson χ 2 test or Fisher’s exact test for categorical variables, depending on the expected count of events in a 2 × 2 contingency table. Comparisons were made between SSRI users and SSRI non-users in terms of demographic systemic conditions and other factors; odds ratios (OR) and their 95% confidence intervals (CI) were computed.

An implant-level model with the implant as the statistical unit was performed in order to assess the effects of age, sex, implant length, implant diameter, implant surface, implant location, and bone augmentation on the failure of implants in patients taking SSRIs. A multivariate generalized estimating equation (GEE) method was used to account for the fact that repeated observations (several implants) were available for a single patient. All models were adjusted for clustering of subject and implants in a binary logistic regression model using GEE with a binomial distribution and a logit link function, while assuming an exchangeable working correlation structure, to assess the relationship between implant failure (dependent variable) and the risk factors (independent variables). A Wald χ 2 test was used to analyze the statistical significance of each parameter within the model.

Furthermore, a multilevel mixed-effects parametric survival analysis was used – a patient-based multilevel analysis – to assess the association between SSRI intake and dental implant failure, accounting for the fact that repeated observations (several implants) were available for a single patient (cluster effect). Because there was little prior knowledge about the appropriate shape of survival probability, parametric frailty models including five different parametric models (Weibull, exponential, log-logistic, log-normal, and gamma) were extended to allow any number of normally distributed random effects. The Akaike information criterion (AIC) was used to choose the best fit survival model. The intake of SSRIs was the exposure variable, and all analyses were adjusted for the following potential confounders: age, sex, implant length, implant diameter, implant surface, implant location, and bone augmentation.

In order to verify multicollinearity, a correlation matrix of all of the predictor variables with a significant OR ( P -value cut-off point of 0.1) was scanned to determine whether there were some high correlations among the predictors. Collinearity statistics obtaining the variance inflation factor (VIF) and tolerance statistics were also performed to detect more subtle forms of multicollinearity. The results of the final multivariate GEE model were presented as the estimated OR, and the results of the multilevel mixed-effects parametric survival analysis were presented as the estimated hazard ratio (HR) of each significant prognostic variable ( P < 0.05).

Kaplan–Meier survival curves were plotted to describe the cumulative proportion of dental implant failure stratified by use of SSRIs (yes/no), and a comparison among groups was analyzed by log-rank test.

The level of statistical significance was set at P < 0.05. All data were analyzed statistically using the software Stata version 14 (StataCorp LP, College Station, TX, USA) and IBM SPSS Statistics version 23 (IBM Corp., Armonk, NY, USA). The study was approved by the Regional Ethics Review Board in Lund and was performed in accordance with the STROBE guidelines for observational studies.


Following the application of the eligibility criteria, a total of 931 implants placed in 300 patients were included in the study; 35 implant failures were reported. Of these 931 implants, 460 were placed in 145 men (mean age 55.9 ± 18.5, range 15.9–82.6 years) and 471 were placed in 155 women (mean age 56.0 ± 17.8, range 14.9–90.8 years; P = 0.665, Mann–Whitney test). A total of 48 implants were placed in 18 SSRI users, whereas 883 implants were placed in 282 non-users. The number of implants per patient ranged from 1 to 15 (mean 3.10 ± 2.49). The mean follow-up time was 2741 ± 2258 days (1578 ± 1717 days for SSRI users, 2804 ± 2267 days for non-users; P < 0.000, Mann–Whitney test).

The implant failure rate was 12.5% (6/48) for SSRI users and 3.3% (29/883) for non-users ( P = 0.007, Fisher’s exact test). The 35 failed implants were lost at a mean of 756 ± 958 days (range 24–3980 days). Eleven of the 35 failed implants (31.4%) were lost prior to the abutment connection procedure or second stage surgery (one in SSRI users, 10 in non-users). Eighteen of the 35 failed implants (51.4%) were lost up to 1 year after surgery.

All implants were inserted with open flap surgery. Nine implants were immediately loaded and three implants were placed in fresh extraction sockets, all in SSRI non-users. Thirty-five implants were non-submerged (five in SSRI users, 30 in non-users). The abutment connection surgery was performed after a mean healing time of 159 ± 63 days for the SSRI users and 150 ± 59 days for the non-users ( P = 0.012, Mann–Whitney test). All 469 ‘turned/machined’ implants were Brånemark implants (Nobel Biocare AB, Göteborg, Sweden). The ‘enlarged-surface’ implants were mostly Brånemark MKIII implants with a TiUnite surface ( n = 386); the rest ( n = 76) were Astra TiOblast, OsseoSpeed, XiVE, Frialit-2 (Dentsply Implants, Mölndal, Sweden) and Straumann SLA (Straumann, Basel, Switzerland) implants. There were no hydroxyapatite-coated surface implants in this group.

Comparisons between users and non-users of SSRIs in terms of demographic systemic conditions and other factors ( Table 1 ) showed that people from the age range >30 to ≤60 years and those >60 years were more likely to be taking SSRIs than younger patients (≤30 years), although this did not reach statistical significance. More women were taking SSRIs than men. Furthermore, fewer implants were placed in the mandible in relation to the anterior maxilla, and fewer patients were followed up for longer periods in the SSRI user group than in the non-user group, and these differences were statistically significant. The groups of SSRI users and non-users were comparable in terms of the other factors.

Table 1
Description of the cohort by implants ( n = 931) between users and non-users of SSRIs.
Variables SSRIs OR (95% CI) P -value
Users, n (%) Non-users, n (%)
Age (years)
≤30 3 (6.3) 127 (14.4) 1
>30 to ≤60 21 (43.8) 311 (35.2) 2.859 (0.838–9.752) 0.093
>60 24 (50.0) 445 (50.4) 2.283 (0.677–7.705) 0.183
Male 3 (6.3) 457 (51.8) 1
Female 45 (93.8) 426 (48.2) 16.092 (4.964–52.166) <0.001
Implant length (mm)
6.0–10.0 11 (22.9) 179 (20.3) 1
10.5–14.0 27 (56.3) 425 (48.1) 1.034 (0.502–2.129) 0.928
15.0–20.0 10 (20.8) 279 (31.6) 0.583 (0.243–1.402) 0.228
Implant diameter (mm)
3.00–3.50 6 (12.5) 85 (9.6) 1
3.70–4.10 42 (87.5) 774 (87.7) 0.769 (0.318–1.861) 0.560
4.20–5.00 0 (0) 24 (2.7)
Implant surface
Turned 22 (45.8) 447 (50.6) 1
Enlarged 26 (54.2) 436 (49.4) 1.212 (0.676–2.170) 0.519
Implant location
Anterior maxilla 27 (56.3) 308 (34.9) 1
Posterior maxilla 9 (18.8) 158 (17.9) 0.650 (0.298–1.415) 0.278
Anterior mandible 6 (12.5) 235 (26.6) 0.291 (0.118–0.717) 0.007
Posterior mandible 6 (12.5) 182 (20.6) 0.376 (0.152–0.928) 0.034
Bone augmentation
No 48 (100) 852 (96.5) 1
Yes 0 (0) 31 (3.5)
Follow-up time (years)
≤1 7 (14.6) 91 (10.3) 1
>1 to ≤5 32 (66.7) 307 (34.8) 1.355 (0.579–3.172) 0.484
>5 to ≤10 4 (8.3) 207 (23.4) 0.251 (0.072–0.879) 0.031
>10 5 (10.4) 278 (31.5) 0.234 (0.072–0.755) 0.015
SSRIs, selective serotonin reuptake inhibitors; OR, odds ratio; CI, confidence interval.

The multivariate GEE model ( Table 2 ) showed that the intake of SSRIs did not exert a statistically significant influence on implant failure, nor did age, sex, implant diameter, or bone augmentation. However, longer implants failed less in comparison to shorter ones, as did enlarged-surface implants in relation to turned implants and implants placed in the anterior mandible in relation to the anterior maxilla. Furthermore, implants failed less in those with longer follow-ups in comparison to shorter periods of observation.

Table 2
Multivariate generalized estimating equation (GEE) logistic regression model (total implants, n = 931).
Variables Failed/survived, n/n OR (95% CI) P -value
No 29/854 1
Yes 6/42 2.316 (0.168–31.932) 0.530
Age (years)
≤30 4/126 1
>30 to ≤60 23/309 1.936 (0.318–11.781) 0.474
>60 8/461 0.224 (0.026–1.916) 0.172
Male 10/450 1
Female 25/446 1.846 (0.655–5.203) 0.246
Implant length (mm)
6.0–10.0 18/172 1
10.5–14.0 12/440 0.203 (0.077–0.536) 0.001
15.0–20.0 5/284 0.172 (0.036–0.827) 0.028
Implant diameter (mm)
3.00–3.50 3/88 1
3.70–4.10 31/785 0.880 (0.124–6.222) 0.898
4.20–5.00 1/23 1.665 (0.113–24.545) 0.710
Implant surface
Turned 21/448 1
Enlarged 14/448 0.085 (0.015–0.495) 0.006
Implant location
Anterior maxilla 15/320 1
Posterior maxilla 6/161 0.341 (0.097–1.197) 0.093
Anterior mandible 3/238 0.124 (0.019–0.805) 0.029
Posterior mandible 11/177 0.490 (0.101–2.386) 0.377
Bone augmentation
No 32/868 1
Yes 3/28 3.220 (0.698–14.847) 0.134
Follow-up time (years)
≤1 18/80 1
>1 to ≤5 11/328 0.088 (0.026–0.295) <0.001
>5 to ≤10 5/206 0.069 (0.012–0.415) 0.004
>10 1/282 0.002 (0.000–0.047) <0.001
SSRIs, selective serotonin reuptake inhibitors; OR, odds ratio; CI, confidence interval.

The multilevel mixed-effects parametric survival analysis ( Table 3 ) adjusted for the aforementioned confounders using a Weibull model based on the AIC selection, showed that the intake of SSRIs did not significantly affect the implant survival rate, and nor did age, sex, implant diameter, implant surface, and implant location. Longer implants had a statistically significant HR, as did bone augmentation procedures.

Dec 14, 2017 | Posted by in Oral and Maxillofacial Surgery | Comments Off on Is the intake of selective serotonin reuptake inhibitors associated with an increased risk of dental implant failure?
Premium Wordpress Themes by UFO Themes