Study Selection and Characteristics
Among the records identified in the preliminary search, an upward trend has been observed in the number of records published from 1986 to the present. A decadal view (see Fig. 1 – upper side) shows the temporal evolution of the number of publications. Figure 1 (lower side) depicts the most frequently used keywords.
The systematic search initially identified a total number of 1.726 records. After removing duplicates, 1.286 records were screened through the reading of titles, abstracts and keywords. A total of 1.137 studies were excluded for different reasons, mainly because they focused on animal models, did not evaluate PPI, or were systematic reviews and/or meta-analyses. Thus, 149 studies were selected for full-text retrieval and subsequent eligibility assessment. Of these, 39 records met the inclusion criteria. Two studies that were identified through the snowballing approach were also included, reaching 41 studies. Figure 2 shows a detailed flowchart of the process and the causes of exclusion.
Figure 2 PRISMA flowchart. Notes: 1. PPI: prepulse inhibition; n: number of records; OS: object of study; OCD: obsessive-compulsive disorder; 2. From Page et al. [60]. Distributed under the terms of the Creative Commons Attribution License
Bibliometric Analysis
199 authors co-authored the selected records. The authors with more articles were C. Grillon, S. M. Southwick, M. A. Geyer, F. X. Vollenweider, and M. Davis. The co-authorship analysis (see Fig. 3) shows four major groups connected by C. Grillon, S. M. Southwick, and A. M. Rasmusson. The most cited authors were C. Grillon (631 citations, 126 citations per paper), M. Geyer (604 citations, 201 citations per paper), and D. I. Braff (594 citations, 297 citations per paper).
The articles were published in 32 journals. Specifically, the journals publishing most articles on this topic were Journal of Affective Disorders (5 papers), and Biological Psychiatry (3 papers). With respect to citations, articles published in Biological Psychiatry received a total of 483 citations (3 papers), followed by the Journal of Neurology, Neurosurgery and Psychiatry (382 citations in one paper). Figure 4 shows a network data map with a detailed analysis of journals and citations.
General Data Description
A summary of the data extracted from the studies is presented in Table 2. The overall mean sample size was 115 participants (SD = 18.7), with a smaller sample size in the patient groups than in the controls (Mean = 42.32, SD = 56.67, and Mean = 46.1, SD = 59.15, respectively). The mean age of the patients was 39.2 years (SD = 14.67), and 37.49 (SD = 14.32) for the control groups. As for the gender, there were, on average, fewer men than women in both the control (47.39%, and 52.61%, respectively), and patient groups (44%, and 56%, respectively).
The psychopathological scales more used were the structured clinical interviews of the DSM-III (9.7%) and DSM-IV (29.2%), and the diagnostic criteria of the International Classification of Diseases (7.31%). Other psychometric scales used were the Clinician-Administered PTSD Scale (9.7%) and the State-Trait Anxiety Test (4.8%). Most of the studies reported the use of psychopharmacology in the pathological sample (68.29%, 28/41), being antidepressants (19.5%) and anxiolytics (9.7%) the most common.
Eighteen studies included patients diagnosed with trauma-, stress- and anxiety-related disorders (44%), sixteen included mood disorders (39%), and seven evaluated patients with neurocognitive disorders (17%). All studies had a case-control design evaluating startle response and PPI, including participants with some psychopathology (97.56%), while the control groups were composed of healthy subjects. In three studies [67, 68, 94], participants were exposed to an experimental induction of emotions, but the control group did not receive such induction. A summary of the main sociodemographic data of the participants, and the experimental variables appears in Table 2.
Table 2
Sociodemographic data and experimental conditions of included studies
Study | N total (% female) | Mean age (SD) | Diagnosis | ISI (ms) | Startle stimuli / Prepulse stimuli |
Swerdlow et al. [64] | 44 (32%) | 45 (2.9) | HD | 30, 60, 120 | 116dB (40ms) / 85dB (20ms) |
Grillon et al. [65] | 48 (0%) | 41.3 (5) | PTSD | 120 | 98-103dB (40ms) / 70dB (30ms) |
Grillon et al. [66] | 66 (53%) | 13.6 (NR) | Anxious sensitivity | 120, 4000 | 106dB (40ms) / 70dB (30ms) |
Grillon et al. [67] | 65 (0%) | 44.2 (4) | PTSD | 120 | 93-103dB (40ms) / 70dB (30ms) |
Grillon et al. [68] | 52 (0%) | 48 (3.7) | PTSD | 120 | 103dB (40ms) / 70dB (30ms) |
Perry et al. [69] | 48 (38%) | 34.6 (11.7) | BD | 30, 60, 120 | 115dB (40ms) / 86dB (20ms) |
Ludewig et al. [70] | 41 (49%) | 35.9 (10) | PD | 30, 60, 120, 240 | 115dB (40ms) / 86dB (20ms) |
Muñoz et al. [71] | 32 (NR) | 47.4 (10.4) | HD | 50, 70, 100, 150 | A: 90dB / 80dB; T: 3xU / 1.5xU |
Hejl et al. [72] | 97 (59%) | 72.6 (5.3) | AD, MCI | 30, 60, 120 | 115dB (40ms) / 85dB (40ms) |
Perry et al. [73] | 46 (52%) | 34 (9.3) | MDD | 30, 60, 120 | 118dB (40ms) / 72-86dB (20ms) |
Perriol et al. [74] | 40 (NR) | 73.3 (NR) | PD, AD | 60, 120, 3000 | 115dB (40ms) / 80dB (40ms) |
Rich et al. [75] | 29 (55%) | 12.9 (2.5) | BD | 60, 120 | 104dB (50ms) / 70dB (50ms) |
Barrett et al. [76] | 43 (49%) | 44.3 (13.2) | BD | 60, 120 | 111dB (40ms) / 75-85dB (40ms) |
Ludewig et al. [77] | 42 (50%) | 34.5 (10) | PD | 30, 60, 120, 240 | 115dB (40ms) / 86dB (20ms) |
Lipschitz et al. [78] | 51 (100%) | 16.5 (2.8) | PTSD | 120, 2000 | 104dB (40ms) / 72dB (40ms) |
Quednow et al. [79] | 38 (42%) | 35.2 (10.8) | MDD, Dysthymia | 140 | 116dB (40ms) / 86dB (20ms) |
Ueki et al. [80] | 70 (69%) | 70.2 (8.6) | AD, MCI | 50 | 115dB (50ms) / 85dB (30ms) |
Carroll et al. [81] | 67 (51%) | 34.5 (8.2) | BD | 120 | 95dB (50ms) / 65dB (50ms) |
Giakoumaki et al. [82] | 57 (NR) | 32 (7.2) | BD | 60, 120 | 115dB (40ms) / 85dB (20ms) |
Duley et al. [83] | 38 (61%) | 21.2 (0.4) | Trait anxiety | 30, 60, 120 | 102dB (40ms) / 70dB (40ms) |
Gogos et al. [84] | 61 (51%) | 41.7 (11.3) | BD | 60, 120 | 115dB (40ms) / 74-86dB (20ms) |
Holstein et al. [85] | 51 (76%) | 38.7 (2.2) | PTSD | 60, 120, 2000 | 115dB (40ms) / 86dB (20ms) |
McMillan et al. [86] | 50 (76%) | 22.9 (5.8) | Anxious sensitivity | 120 | 105dB (50ms) / 70dB (25ms) |
Vrana et al. [87] | 100 (50%) | 42 (10.8) | PTSD | 60, 120, 240 | 100dB (50ms) / 70dB (20ms) |
Zoetmulder et al. [88] | 82 (49%) | 59.7 (9.1) | PD | 30, 60, 120, 300 | 115dB (40ms) / 75-85dB (20ms) |
Ivleva et al. [89] | 214 (56%) | 34.3 (11.7) | BD | 120, 4500 | 116dB (40ms) / 80dB (20ms) |
Comasco et al. [90] | 204 (100%) | 30.4 (4.9) | Trait anxiety | 100 | 115dB (40ms) / 72-86dB (20ms) |
Vrana et al. [91] | 95 (53%) | 42 (10.8) | PTSD | 60, 120, 240 | 100dB (50ms) / 70dB (20ms) |
Sánchez-Morla et al. [92] | 102 (56%) | 40.9 (10.5) | BD | 60, 120 | 118dB (40ms) / 80dB (20ms) |
Pineles et al. [93] | 47 (100%) | 31.9 (9.2) | PTSD | 120 | 100dB (50ms) / 70dB (20ms) |
De la Casa et al. [94] | 22 (64%) | 21 (NR) | Stress induction | 40, 60, 80 | 95dB (20ms) / 75dB (50ms) |
Comasco et al. [95] | 170 (100%) | 30.9 (4.8) | Postpartum MDD | 100 | 115dB (40ms) / 72-86dB (20ms) |
Echiverri-Cohen et al. [96] | 67 (67%) | 32.7 (13.4) | PTSD | 30, 60, 120 | 105dB (50ms) / 75dB (25ms) |
Matsuo et al. [97] | 471 (53%) | 39.9 (11.7) | MDD | 60, 120 | 115dB (40ms) / 86-90dB (20ms) |
Millian-Morell et al. [98] | 87 (48%) | 70.3 (11.9) | PD | 60, 120, 1000 | 115dB (40ms) / 85dB (20ms) |
Meteran et al. [99] | 45 (47%) | 45.9 (13.1) | PTSD | 60, 120 | 115dB (20ms) / 76-85dB (20ms) |
Bo et al. [100] | 63 (40%) | 26.3 (6.7) | BD | 120 | 100dB (40ms) / 65dB (150ms) |
Matsuo et al. [101] | 338 (59%) | 40.4 (11.1) | BD | 60, 120 | 115dB (NR) / 86-90dB (NR) |
Massa et al. [102] | 1143 (59%) | 38.6 (14) | BD | 120 | 116dB (40ms) / 90dB (20ms) |
Storozheva et al. [103] | 240 (47%) | 33.2 (1.2) | GAD | 60, 120, 2500 | 110dB (40ms) / 85dB (20ms) |
San-Martin et al. [104] | 44 (39%) | 26.6 (7.7) | BD | 30, 60, 120 | 115dB (40ms) / 85dB (20ms) |
Note: Abbreviations: AD, Alzheimer’s disease; BD, bipolar disorder; dB, decibels; GAD, general anxiety disorder; HD, Huntington’s disease; ISI, interstimulus interval; MCI, mild cognitive impairment; MDD, major depressive disorder; ms, milliseconds; N, number of participants; NR, not reported; PD, panic disorder; PD, Parkinson’s disease; PTSD, post-traumatic stress disorder; SD, standard deviation.
Experimental Conditions
The experimental conditions are summarized in Table 2. Most of the studies used more than one interstimulus interval between the prepulse and pulse (70.7%, 29/41), with 120 ms (85.4%), 60 ms (61%), and 30 ms (24.4%) being the most common intervals. The overall mean pulse intensity was 109.41 dB (SD = 7.69; range 90–118), with a mean duration of 41 ms (SD = 6.4; range 20–50). The mean prepulse duration was 29.74 ms (SD = 22.03; range 22–150), with a mean intensity of 78.98 dB (SD = 6.91; range 65–90). All experiments registered the electromyographic response of the orbicularis oculi muscle as the measure of the startle response. Two of the studies recorded it bilaterally [83, 103], while the remaining studies recorded it on the right (68.3%, 28/41) or left orbicularis muscle (26.8%, 11/41). All studies used an acoustic sensory modality, with two studies using also tactile stimuli [64, 71].
Trauma-, Stress-, And Anxiety-related Disorders
The summary of the main findings from the studies on trauma-, stress-, and anxiety-related disorders appears in Table 3. The studies included PTSD (n = 10, 55.6%), panic disorder (n = 2, 11.1%), trait anxiety (n = 2, 11.1%), anxious vulnerability (n = 2, 11.1%), and generalized anxiety disorder (n = 1, 5.6%) patients. One study exposed subjects to an experimental induction of stress (5.6%).
Regarding startle response intensity, most studies found a higher magnitude of startle response for the patients than the control group (55.5%), whereas fewer studies reported a lower magnitude (16.7%). Four studies did not find differences between groups [65, 70, 78, 96] and one study did not report data on startle magnitude [83]. More specifically, an increased startle response appeared for PTSD [67, 68, 85, 87, 91, 99] and for anxious vulnerability patients [66, 86].
PPI results were more consistent, with 14 of the 18 studies in this group reporting disrupted PPI in the pathological group compared to the control group (77.8%). The studies without differences in PPI only included patients with PTSD [68, 78, 85, 99].
Table 3
Summary of studies on PPI deficit that compared trauma-, stress-, and anxiety-related disorders patients to matched controls
Study | Startle response (SR) | Prepulse inhibition (PPI) | Neurobiology proposed | Cognition proposed |
Storozheva et al. [103] | GAD > Control: ↓ SR mg. ↑ SR lat. | GAD > Control: ↓ %PPI (ISI 60ms) | GAD > Control: ↑ LH RT ↓ SR ↓ PFC RT ↓ PPI | Misinterpretation of contextual cues RT ↓ PPI |
Grillon et al. [65] | No differences | PTSD > Control: ↓ %PPI | NR | Affective flattening, avoidance and re-experiencing RT ↓ PPI |
Grillon et al. [67] | S1: No differences S2: PTSD > Control: ↑ SR | PTSD > Control: ↓ %PPI | ↑ Activity of the NST and hippocampus RT ↓ PPI | Re-experiencing RT ↑ SR mg. |
Grillon et al. [68] | PTSD > Control: ↑ SR mg. | No differences | ↑ Activity of the NST and hippocampus RT ↑ SR mg. | Re-experiencing & avoidance RT ↑ SR mg. |
Lipschitz et al. [78] | No differences | No differences | NR | NR |
Holstein et al. [85] | PTSD > Control: ↑ SR mg. | No differences | NR | NR |
Vrana et al. [87] | PTSD > Control: ↓ SR lat. ↑ SR mg. | PTSD > Control: ↓ %PPI (ISI 60 y 120ms) | NR | Hypervigilance RT SR ↓ lat. & ↑ mg. |
Vrana et al. [91] | PTSD > Control: ↑ SR mg. | PTSD > Control: ↓ %PPI | PFC deficit RT ↓ PPI | ↑ Abstinence & planning RT ↑ PPI |
Pineles et al. [93] | PTSD > Control: ↓ SR mg. | PTSD > Control: ↓ %PPI | NR | Re-experiencing & avoidance RT ↓ PPI |
Echiverri-Cohen et al. [96] | No differences | PTSD > Control: ↓ %PPI (ISI 30 y 60ms) | NR | Re-experiencing & avoidance RT ↓ PPI |
Meteran et al. [99] | PTSD > Control: ↑ SR mg. | No differences | PFC deficit RT ↑ SR | Hallucinations & other psychotic symptoms RT ↑ SR |
De la Casa et al. [94] | Stress > Control: ↓ SR mg. | Stress > Control: ↓ %PPI (ISI 60 y 80ms) | ↑ Dopaminergic activity RT stress induction RT ↓ PPI | NR |
Duley et al. [83] | NR | TA > Control: ↓ %PPI | NR | Exercise modulates PPI deficit in anxiety |
Comasco et al. [90] | PW > Control: ↑ SR mg. | PW > Control: ↓ %PPI PW + TA > PW: ↓ %PPI PW + TA + SSRI > PW + TA: ↓ %PPI | ↑ Estrogens & catecholaminergic genotype RT ↓ PPI | Attention and executive deficits RT ↓ PPI |
Grillon et al. [66] | AS > Control: ↑ SR mg. | AS > Control: ↓ %PPI | NR | NR |
McMillan et al. [86] | AS > Control: ↑ SR mg. | AS > Control: ↓ %PPI | NR | Difficulty disengaging attention RT ↓ PPI |
Ludewig et al. [70] | No differences | PD > Control: ↓ %PPI PD + A > PD-A: ↓ %PPI | NR | Trait anxiety RT ↓ PPI |
Ludewig et al. [77] | PDnM > Control: ↑ SR mg. | PDnM > Control: ↓ %PPI PDnM > PDM: ↓ %PPI (240ms) | ↑ Dopaminergic & amygdala activity RT ↓ PPI | Deficit in interpreting somatic symptoms RT ↑ SR |
Note: Abbreviations: ↑, increase or hyperactivation of; ↓, deficit or hypoactivation of; AS, anxious sensitivity; GAD, general anxiety disorder; ISI, interstimulus interval; Lat., latency; Mg., magnitude; ms, milliseconds; NR, not reported; NST, nucleus of the stria terminalis; PD, panic disorder; PD + A/-A, panic disorders patients with or without anxiety; PDM/nM, medicated or unmedicated panic disorder patients; PFC, prefrontal cortex; PTSD, post-traumatic stress disorder; PW, pregnant women; PW + TA, pregnant women with trait anxiety; PW + TA + SSRI, pregnant women with trait anxiety medicated with selective serotonin reuptake inhibitors; RH, left hemisphere; RT, related to; S1/2, first and second sessions; PPI, prepulse inhibition; SR, startle response; TA, trait anxiety.
Mood Disorders
A detailed analysis of all the variables for each study included in this group of disorders is presented in Table 4. The studies included bipolar disorder (n = 12, 75%), major depressive disorder (n = 4, 25%), and dysthymia (n = 1, 6.25%).
Regarding startle response intensity, most studies did not find differences between the groups (68.75%). However, three studies reported longer latencies of the startle response in the groups of patients [81, 92, 102].
Nine studies (56.25%) revealed reduced PPI in patients compared to the control group. Specifically, 7 of 12 studies with bipolar disorder patients informed of a PPI deficit [69, 82, 84, 92, 100–101, 104]. Only one study reported reduced PPI in women from the control group compared to bipolar patients [84].
The results from studies with major depressive disorder patients were quite contradictory. Thus, two of the studies found disrupted PPI for major depressive disorder patients [95, 97], and two studies did not find differences between groups [73, 79].
Table 4
Summary of studies on PPI deficit that compared mood-related disorders patients to matched controls
Study | Startle response (SR) | Prepulse inhibition (PPI) | Neurobiology proposed | Cognition proposed |
Perry et al. [73] | No differences | No differences (tendency: MDD > Control: ↓%PPI) | CSPP deficit RT ↓ PPI | NR |
Quednow et al. [79] | No differences | No differences | NR | Suicide attempt no RT PPI |
Comasco et al. [95] | MDD > Control: ↑ SR mg. | MDD > Control: ↓ %PPI | Genetic risk RT ↓ PPI | Depression & insomnia RT ↓ PPI |
Matsuo et al. [97] | No differences | MDD♂>Control♂: ↓ %PPI | Sexual dimorphism RT PPI | Depression RT ↓ PPI |
Perry et al. [69] | No differences | BD > Controles: ↓ %PPI | CSPP deficit RT ↓ PPI | Cognitive fragmentation RT ↓ PPI |
Rich et al. [75] | No differences | No differences | NR | ADHD symptoms no RT PPI |
Barrett et al. [76] | No differences | No differences | NR | NR |
Carroll et al. [81] | BD > Control: SR ↓mg. ↑lat. | No differences | NR | Depression RT ↑ SR |
Giakoumaki et al. [82] | No differences | BD > grBD > Control: ↓ %PPI | Genetic risk & PFC deficit RT ↓ PPI | ↓ Inhibitory control RT ↓ PPI |
Gogos et al. [84] | BD♂>Control♂: ↓ SR mg. | BD♂>Control♂: ↓%PPI (60ms) BD♀>Control♀: ↑%PPI (120ms) | Increase in 5-HT receptors RT ↑ PPI (♀) | ↑ ISI (120ms) mobilizes attentional resources |
Ivleva et al. [89] | No differences | No differences | NR | NR |
Sánchez-Morla et al. [92] | BD > Control: ↑ SR lat. | BD > Control: ↓ %PPI (ISI 60 y 120ms) | Amygdala déficit RT ↓ PPI | ↓ Social cognition RT ↓ PPI |
Bo et al. [100] | No differences | BD > Control: ↓ %PPI | PFC deficit RT ↓ PPI | ↓ Inhibitory control RT ↓ PPI |
Matsuo et al. [101] | No differences | BD♂>Control♂: ↓ %PPI | NR | Depression RT ↓ PPI |
Massa et al. [102] | BD > Control: ↑ SR lat. | No differences | Genetic risk RT ↓ PPI | ↓ Memory, executive function & SIP RT ↓ SR |
San-Martin et al. [104] | No differences | BD > Control: ↓%PPI (ISI 60ms) | NR | NR |
Note: Abbreviations: ↑, increase or hyperactivation of; ↓, deficit or hypoactivation of; BD, bipolar disorder; CSPP, corticostriatal-pallidopontine limbic circuit; ISI, interstimulus interval; Lat., latency; MDD, major depressive disorder; Mg., magnitude; ms, milliseconds; NR, not reported; PFC, prefrontal cortex; PPI, prepulse inhibition; RT, related to; SIP, speed of information processing; SR, startle response.
Neurocognitive Disorders
This group of studies included patients with Parkinson’s (n = 3, 42.8%), Alzheimer's (n = 3, 42.8%) and Huntington’s (n = 2, 28.6%) diseases. Two additional studies included patients with mild cognitive impairment. A detailed analysis of all the variables for each study included in this group of disorders is depicted in Table 5.
Regarding startle response, 3 studies (42.8%) reported longer latencies for the patients, and 3 (42.8%) did not find differences between groups. As for PPI results, 71.4% of the studies reported reduced PPI in the group of patients compared to the control group, and only one study did not find any differences [72]. Studies including patients with movement disorders showed the most consistent results. Specifically, all studies including patients with Huntington’s disease found reduced PPI [64, 71], with a greater deficit in patients with chorea [71]. Consistency was also high for Parkinson’s disease patients, showing a generalized PPI deficit [74, 88], except for one study by Millian-Morell et al. [98] which reported an increase in PPI.
Regarding Alzheimer's type dementia, two studies found reduced PPI [74, 80], and one study did not find significant differences [72]. Patients with mild cognitive impairment either did not differ from [72] or showed increased PPI compared to controls [80].
Table 5
Summary of studies on PPI deficit that compared neurocognitive disorders patients to matched controls
Study | Startle response (SR) | Prepulse inhibition (PPI) | Neurobiology proposed | Cognition proposed |
Perriol et al. [74] | NR | PD > AD > Control: ↓ %PPI (ISI 120ms) | Subcortical-thalamo-cortical system dysfunction RT ↓ PPI | Exogenous care RT PPI modulation (ISI 120ms) |
Zoetmulder et al. [88] | No differences | PD > Control: ↓ %PPI (ISI 60 y 120ms) | Striatal dysfunction RT ↓ PPI | NR |
Millian-Morell et al. [98] | No differences (tendency: PD > Control: ↑ SR lat.) | PD > Control: ↑ %PPI (ISI 120ms) | Deficits in basal ganglia, PFC and dopaminergic network RT ↑ PPI | Motor coordination deficit RT ↑ PPI |
Hejl et al. [72] | No differences | No differences | The cholinergic system would have a weak relationship with PPI | NR |
Ueki et al. [80] | No differences | MCI > Control: ↑ %PPI AD > Control: ↓ %PPI | Deficits in entorhinal cortex in early stages of AD RT ↓ PPI | Cognitive-behavioral dementia symptoms RT ↓ PPI |
Swerdlow et al. [64] | HD > Control: ↑ SR lat. | HD > Control: ↓ %PPI | Deficits in GABA efferent circuit from striatum-pale RT ↓ PPI | Inhibitory and executive deficits RT ↑ SR latency |
Muñoz et al. [71] | HD > Control: ↑ SR lat. | HD > Control: ↓ %PPI HD + cm > HD: ↓ %PPI | Glutamatergic dysfunction RT ↓ PPI | NR |
Note: Abbreviations: ↑, increasement or hyperactivation of; ↓, deficit or hypoactivation of; BD, bipolar disorder; AD, Alzheimer’s disease; HD, Huntington’s disease; HD + cr: Huntinton’s disease patients with choreic movements; ISI, interstimulus interval; Lat., latency; MCI, mild cognitive impairment; MDD, major depressive disorder; ms, milliseconds; NR, not reported; PD, Parkinson’s disease; PFC, prefrontal cortex; PPI, prepulse inhibition; RT, related to; SR, startle response.
Assessment Of Risk Of Bias
The results of the methodological quality analysis performed using the Newcastle-Ottawa Scale to assess the risk of bias of the studies included in the review are summarized in Table 6. In general, most of the studies had a low risk of methodological bias. The overall mean quality was 7.19 stars (SD = 0.95; range 6–9). Specifically, in the category of study selection, most of them presented a good definition of the cases, as well as representative samples of the population. The comparability analysis showed that most of the studies controlled for sex and age. Finally, the results of the level of exposure of the participants to the evaluation methods and the experimental paradigm revealed a moderate risk of bias, mainly due to the fact that many studies did not specify the non-response rate or the method of ascertainment for cases and controls.
Table 6
Risk of bias of the studies included in the systematic review assessed by Newcastle-Ottawa Quality Assessment Scale (NOS)
Study | Selection | Comparability | Exposure | | | Risk of bias assessment |
Case definition | Represen-tativeness | Selection of controls | Definition of controls | Confounding factors (*) | Ascertainment of exposure | Same method of ascertainment | Nonresponse rate |
Swerdlow et al. [64] | * | * | * | * | */* | * | * | | 8 |
Grillon et al. [65] | * | * | * | * | */* | * | * | | 8 |
Grillon et al. [66] | | | * | * | */* | * | * | | 6 |
Grillon et al. [67] | * | | * | * | */* | * | * | * | 8 |
Grillon et al. [68] | * | | * | * | */* | * | * | | 7 |
Perry et al. [69] | * | | * | | */* | * | * | | 6 |
Ludewig et al. [70] | * | | * | * | */* | * | * | | 7 |
Muñoz et al. [71] | * | | * | * | */ | * | * | | 6 |
Hejl et al. [72] | * | | * | * | */* | | * | * | 7 |
Perry et al. [73] | * | | * | * | */* | | * | | 6 |
Perriol et al. [74] | * | | * | | */* | * | * | | 6 |
Rich et al. [75] | * | | * | * | */* | * | * | | 7 |
Barrett et al. [76] | * | | | * | */* | * | * | | 6 |
Ludewig et al. [77] | * | * | * | * | */* | * | * | | 8 |
Lipschitz et al. [78] | * | | * | | */* | * | * | | 6 |
Quednow et al. [79] | * | | * | * | */* | | * | * | 7 |
Ueki et al. [80] | * | * | * | * | */* | * | * | | 8 |
Carroll et al. [81] | * | | * | * | */* | | * | * | 7 |
Giakoumaki et al. [82] | * | | * | * | */* | * | * | * | 8 |
Duley et al. [83] | * | | | * | */* | * | * | | 6 |
Gogos et al. [84] | * | | * | * | */* | * | * | | 7 |
Holstein et al. [85] | * | | * | * | */* | * | * | * | 8 |
McMillan et al. [86] | * | | * | * | */* | | * | | 6 |
Vrana et al. [87] | * | | * | * | */* | * | * | | 7 |
Zoetmulder et al. [88] | * | | * | * | */* | | * | * | 7 |
Ivleva et al. [89] | * | * | * | * | */* | | * | * | 8 |
Comasco et al. [90] | * | * | * | * | */* | * | * | * | 9 |
Vrana et al. [91] | * | | * | * | */ | * | * | | 6 |
Sánchez-Morla et al. [92] | * | | * | * | */* | | * | | 6 |
Pineles et al. [93] | * | | * | * | */* | | * | * | 7 |
De la Casa et al. [94] | * | | * | * | */* | | * | * | 7 |
Comasco et al. [95] | * | * | * | * | */* | * | * | | 8 |
Echiverri-Cohen et al. [96] | * | | * | * | */* | * | * | | 7 |
Matsuo et al. [97] | * | * | * | * | */* | * | * | * | 9 |
Millian-Morell et al. [98] | * | | * | * | */* | | * | * | 7 |
Meteran et al. [99] | * | | * | * | */* | * | * | * | 8 |
Bo et al. [100] | * | | * | * | */* | * | * | | 7 |
Matsuo et al. [101] | * | * | * | * | */* | * | * | * | 9 |
Massa et al. [102] | * | * | * | * | */* | * | * | * | 9 |
Storozheva et al. [103] | * | * | | * | */* | * | * | * | 8 |
San-Martin et al. [104] | * | | * | * | */* | * | * | | 7 |
Notes
1. * means a point in the category in which it is indicated; 2. Principal cofounding factors were sex, age and smoking status