Explore the vast range of titles available.

The utilisation of marine current turbines with horizontal axis for electrical power production offers a sustainable and clean option to augment traditional energy technologies and enhance the expansion of renewable energies in the marine field. Some prototypes of these turbines are now available and operating. The advantages of tidal currents for renewable energies when compared with wind resources are that sea water is about 800 times denser that air (and the extracted energy is proportional to density) and the nature of currents that results in a more predictable resource than wind resources. But these advantages set certain limitations and technical problems, some of them related with the hydrodynamic design of rotor blades since the design of a marine turbine is some kind different of the techniques used for wind turbines. Methodologies need be established to describe the physical and operational performance of the turbines, allowing their design to be investigated and performance evaluated. This paper describes the hydrodynamics of rotor design that is projected in order to extract the maximum power from the tide. The blade element momentum theory is used for the rotor modelling, different aspects and limitations of this approach when applied to marine rotor design are commented, and examples are presented.

Objective: To evaluate the short term effect of air pollution on cardiovascular admissions in 14 Spanish cities Methods: The period under study was from 1995 to 1999. Daily emergency admissions for all cardiovascular diseases (CVD) and heart diseases (HD) were obtained from hospital records, and the corresponding daily levels of particulates, SO2, NO2, CO, and ozone were recorded. The magnitude of association was estimated using Poisson generalised additive models controlling for confounding and overdispersion. For each cause, lagged effects, up to three days, of each pollutant were examined and combined estimates were obtained. For ozone the analyses were restricted to the warm period. One and two pollutant models were performed. Results: Associations were more consistent in lag 0 (concurrent day) and 1 (lag 0–1), except in the case of ozone where there was a more delayed relation (lag 2–3). For combined estimates an increase of 10 μg/m3 in the PM10 levels in lag 0–1 was associated with an increase of 0.9% (95% CI: 0.4 to 1.5%) in the number of hospital admissions for CVD, and 1.6% (0.8 to 2.3%) for HD. For ozone the corresponding estimates for lag 2–3 were 0.7% (0.3 to 1.0) for CVD, and 0.7% (0.1 to 1.2) for HD. An increase of 1 mg/m3 in CO levels was associated with an increase of 2.1% (0.7 to 3.5%) in CVD admissions, and 4.2% (1.3 to 7.1%) in HD admissions. SO2 and NO2 estimates were more sensitive in two pollutant models Conclusions: A short term association between increases in daily levels of air pollutants and the number of daily admissions for cardiovascular diseases, with specificity for heart diseases, has been described in Spanish cities.

BackgroundThe anti-TNF therapy has been demonstrated effective in patients with Spondylarthritis (SpA). The lack of efficacy and adverse events (1) have been reported as the most common reasons for the discontinuation of these drugs. In Rheumatoid Arthritis, the immunogenicity has been proved to be an influential factor in the survival of treatment, but this has not been proved in SpA.ObjectivesTo assess the survival of 3 anti-TNF agents in a cohort of patients with SpA, and to determine if there are differences in the survival of the 3 anti-TNF between patients who develop anti-TNF antibodies and those who do not.MethodsPatients with SpA who started Adalimumab (ADA), Infliximab (IFX) or Etanercept (ETN) as first anti-TNF treatment between 2001–2015 were included. Diagnoses were: Ankylosing Spondylitis (55.5%), Non Radiographic Axial SpA (20%), SpA associated with inflammatory bowel disease (5.9%) and Psoriatic SpA (18.6%). Patients with only peripheral involvement were excluded. Clinical and demographic data were collected, and in the case of ADA and IFX, the presence or absence of anti-drug antibodies were identified. No anti-ETN antibodies were identified. The reasons for drug discontinuation were classified as adverse event, administration reaction, primary inefficacy, secondary inefficacy, recovery and other (this included pregnancy, malignancies, etc). The event was defined as drug discontinuation due to only the first 4 reasons. Cumulative incidence through competitive risk was performed to compare drug survival.ResultsOf the 220 patiens studied, 58.6% were men, with a mean age of 43.5±12.4 years. Out of them, 38.6% (85) started treatment with IFX, 30% (66) with ADA, and 31.4% (69) with ETN. No statistically significant difference (p=0.3) was found in the survival of the 3 anti-TNF: IFX had a probability of drug discontinuation of 32.9% at 5 years, compared to 35.5% of ADA and versus 31.8% of ETN. The most common reasons for discontinuing treatment were: Primary inefficacy for IFX (22.4%), primary and secondary inefficacy for ADA (13.6% each), and others (pregnancy, etc) for ETN (26.1%).30.6% of patients treated with IFX developed anti-IFX antibodies (AIA). Anti-ADA antibodies (AAA) were observed in 22.7% of 66 patients who received ADA. In both drugs, the drug survival was lower in patients who developed anti-TNF antibodies, but this difference was statistically significant only in the group of IFX:Cumulative incidence through competitive risk (probability of drug discontinuation)1 year5 years10 yearsp valueInfliximab NO AIA13.5%25.4%34.3%0.003 WITH AIA11.5%50.0%69.2%Adalimumab NO AAA14.7%33.4%0.5% WITH AAA21.0%43.5%When survival was analyzed by subgroups of patients treated with IFX and ADA who had not developed antibodies compared with ETN, no statistically significant difference was found (p=0.5).ConclusionsIn our cohort, the three anti-TNF showed a similar survival. However, and particularly in the case of IFX, the development of anti-drug antibodies proved to be an influential factor in the survival of treatment.ReferencesRosales-Alexander et al. Drug survival of anti-tumour necrosis factor a therapy in spondyloarthropathies: results from the Spanish emAR II study. Rheumatology (Oxford). 2015 Aug;54(8):1459–63Disclosure of InterestS. García-Carazo Grant/research support from: This work was funded by unrestricted grant from PFIZER, C. Plasencia: None declared, D. Pascual-Salcedo: None declared, D. Peiteado: None declared, G. Bonilla: None declared, L. Nuño: None declared, A. Villalba: None declared, M. Díaz: None declared, F. Arribas: None declared, A. Balsa: None declared

In patients with rheumatoid arthritis (RA), the development of antibodies to infliximab (ATI) is associated with poor clinical response. 1 2-7 Nevertheless, there is no plausible explanation for why not all patients with ATI experience high disease activity. Table 1 Total: 11 patients with RA (%) Gender, female n (%) 11 (100) Age, mean (SD) 54.18±13.24 Autoantibodies: RF positive n (%) 9 (81.8) ACPA positive n (%) 9 (81.8) Disease duration (years), mean (SD) 14.20±6.42 Baseline DAS28, mean (SD) 5.45±1.13 Time under Ifx therapy in years, mean (SD) 8.00±2.75 ATI duration in years, mean (SD) 1.1±0.91 Concomitant treatment MTX alone 3 (27.2) OD 1 (9.0) MTX+OD 6 (54.8) Ifx monotherapy 1 (9.0) ACPA, Anticitrullinated protein antibody; ATI, antibodies to infliximab; DAS28, Disease Activity Score 28; Ifx, infliximab; MTX, Methotrexate; OD, Other DMARDs; RF, Rheumatoid Factor.

Background There is limited evidence regarding longterm effects of tapering of TNF inhibitors (TNFi) in Spondyloarthritis (SpA) patients Objectives To compare longterm clinical outcomes in SpA patients on tapering strategy with SpA patients on standard regimen Methods In this observational study, 117 SpA patients (pts) treated with TNFi [infliximab (Ifx), adalimumab (Ada) and etanercept (Etn)] were included. Two groups were compared: TNFi tapering strategy (Group 1: 74 pts from Spain) vs the control group on standard treatment regimen (Group 2: 43 pts from the Netherlands). A control group for the Ifx was not available. Pts were matched on duration of inactive disease prior to inclusion and duration of followup. All SpA pts had to be at least 6 months in inactive disease (BASDAI<4) to be included. The tapering strategy included dose reduction and/or interval prolongation, in case a flare (BASDAI>4)appeared, the TNFi dose could be increased or the interval could be shortenned to regain low disease activity. The clinical activity was measured, using BASDAI, at different time points: visit-0 (baseline, prior starting the biological therapy), visit-1 (Group1: just before starting tapering; Group 2: after at least 6 months in inactive disease) and visit-2 (the last visit available after visit-1) Results Seventy four pts were included in tapering group (Group 1: 35 with Ifx, 17 with Ada and 22 with Etn) and 43 pts in control group (Group 2: 21 with Ada and 22 with Etn). No statistical differences were seen in the time (years) in inactive disease prior visit-1 (1.23±1.09 in Group 1 vs 0.73±0.18 in Group 2, p=0.243) and followup (years) between visit-1and visit-2 (2.33±1.12 in Group 1 vs 2.40±0.99 in Group 2, p=0.567).No significant differences were found in clinical activity and percentage of pts with flares (see Table 1). The dropout tended to be higher in pts on tapering, being the secondary inefficacy more frequent in this group, but also the remission (see Table 1).The overall drug administered was reduced in tapering group at visit-2 in comparison with the group on standard protocol (dose reduction of 22% in Ifx and an elongation in administration interval of 28.7% in Ifx, 45.2% in Ada and 51.5% in Etn) Table 1 SpA pts (Total=117) Group 1: Group 2: p Tapering group Control group n=74 n=43 BASDAI (m ± SD)  Visit-0 5.82±1.63 5.78±1.30 0.964  Visit-1 1.77±1.02 1.79±0.99 0.724  Visit-2 2.15±1.55 2.11±1.31 0.883 Percentage of pts with flares, n/N (%) 28/74 (37.8%) 11/43 (25.6%) 0.175 Reasons to dropout (number of patients):  Remission 7/74 (9.5%) 0/43 (0%)  Secondary Inefficacy 3/74 (4.1%) 0/43 (0%)  Drop out for other reasons 0/74 (0%) 1/43 (2.3%) At visit-2  Interval administration (weeks), m ± SD Ifx: 11.22±1.80 Ifx: 8.00±0.00 Ada: 3.74±1.21 Ada: 2.00±0.00 <0.001 Etn: 2.09±0.59 Etn: 1.00±0.00 <0.001  Dose administration (mg/kg), m ± SD Ifx: 4.40±0.81 Ifx: 5.00±0.00 Conclusions The tapering strategy in inactive SpA pts results in an important reduction of the drug administered while the disease control remains similar to SpA on standard regimen. However, a small percentage of SpA pts on tapering seems to be more prone to dropout due to secondary inefficacy and also the remission was more often Disclosure of Interest C. Plasencia Grant/research support: Pfizer, G. Wolbink Grant/research support: Pfizer, Speakers bureau: Pfizer, Amgen, E. Kneepkens: None declared, C. Krieckaert Speakers bureau: Abbvie, Pfizer, M. l'Ami: None declared, D. Peiteado: None declared, L. Nuno: None declared, F. Arribas: None declared, M. Nurmohamed Grant/research support: Abbvie, BMS, MSD, Pfizer, UCB, Roche, Consultant for: Abbott, BMS, Pfizer, Roche, Speakers bureau: Abbvie, Pfizer, Roche, E. Martín-Mola Speakers bureau: Pfizer, Abbvie, Roche, UCB, A. Balsa Grant/research support: Pfizer, Speakers bureau: Pfizer, Abbvie, Roche, D. Pascual-Salcedo Grant/research support: Pfizer, Speakers bureau: Pfizer DOI 10.1136/annrheumdis-2014-eular.3208

Background The anti-TNF therapy is an effective treatment in RA patients, however a considerable percentage of patients develop clinical inefficacy (primary or secondary); part of which can be explained by the development of antibodies to infliximab (ATI). In some patients, a dose increase is often used to achieve a clinical improvement, mainly in patients with secondary inefficacy. However, there is no evidence that demonstrates if dose escalation is a good therapeutic option. Objectives To assess whether increasing the infliximab (Ifx) dose is an effective therapeutic option in RA patients who develop clinical inefficacy. Methods The present study included 36 RA patients treated since 2000 with Ifx at La Paz University Hospital, in whom a Ifx dose increase was implemented due to inefficacy [23/36 (64%) primary non-responders and 13/36 (36%) secondary non-responders]. All patients started to receive Ifx at 3 mg/kg i.v. at standard schedule. A therapeutic increase was defined as: Ifx dose higher than 3 mg/kg i.v. (until maximum of 5 mg/kg) or interval time between infusion 6-7 weeks. The clinical activity was measured by the Disease Activity Score 28 (DAS28) at baseline, before starting the dose escalation, at 6 months and at 1 year after the dose increase. The drug and ATI levels were measured by a capture and bridging ELISA, respectively. Statistical analysis was made using SPSS system 11.0. Results Thirty one out of 36 patients (86.1%) were female, with a mean age of 58±13.6 years. The disease duration was 19.2±10.5 years and the time under Ifx therapy was 6.6±3.8 years. Most patients received concomittant therapy with DMARDs [15 (41.7%) methotrexate, 22 (41.7%) other DMARDs and 8 (22.2%) in monotherapy] and 13 (38.2%) received prednisone. All patients were active (DAS28) at baseline and before starting the dose increase (5.4±1.0 at baseline and 4.2±1.2 just before dose escalation). ATI were detected in 13/34 (38.2%) patients at the moment of decision to increase (in 2 patients ATI were already present in the first available sample) and only 2 (5.5%) patients became negative after dose increase. The Ifx increment did not produce significant improvement in clinical activity (DAS28) in RA patients, independently on ATI status [ATI-negative: 4.25±1.4 at pre-increase vs 3.9 ±1.0 at 6 months after increase (p=0.311) and 4.04±1.1 vs 4.03 ±1.3 at 1 year after increase (p=0.970); ATI-positive: 4.09±0.9 at pre-increase vs 4.1±1.3 at 6 months after increase (p=0.945) 3.7 ±0.7 vs 3.0±0.8 at 1 year after increase (p=0.110)]. A total of 26 (76.5%) patients discontinued the Ifx therapy, not identifying differences between ATI-positive patients (15/21, 71.4%) and ATI-negative patients (11/13, 84.6%), p=0.378. Conclusions Infliximab dose escalation does not seem to be an effective therapeutic option in RA patients who develop primary and secondary inefficacy. Further studies are necessary to confirm these results. Disclosure of Interest None Declared

Background In clinical practice of patients with rheumatic diseases, the optimization of the biological therapy (BT) is performed by adjusting drug dose according to clinical activity. To help taking accurate treatment decisions, it will be desirable to have complementary objective tools, next to the ones currently used. Since 2010, at the Rheumatology Department of La Paz University Hospital, serum through levels of TNF inhibitors are taken into account, together with the regular tools to asses clinical activity, in order to improve the therapeutic outcomes. Objectives To analyse the clinical as well as financial impact of the therapeutic drug monitoring (TDM), based on serum trough drug levels, in rheumatoid arthritis (RA) and Spondiloarthritis (SpA) patients in remission or low disease activity. Methods In an observational study of daily clinical practice, a total of 88 patients (43 RA and 45 SpA), treated with three TNF inhibitors [31 with Infliximab (Ifx), 29 with Adalimumab (Ada) and 28 with Etanercept (Etn)] between 2006-2012, were included. All patients were in remission or low clinical activity (DAS28<3.2 ó BASDAI <4) throughout the 7 years analyzed. In each patient two different time periods were examined, pre- and during- TDM practice (1stP: 2006-2009 and 2ndP: 2010-2012) to observe if serum trough drug levels influence on therapeutic decisions as down-titration or cessation of anti-TNF therapy. Drug levels were measured by capture ELISA´s as described (1,2,3). Results The mean timespan of disease duration was 17.52 ±9.38 years and the mean in biological therapy was 5.85 ±1.33 years. All patients were clinically active at the beginning of anti-TNF therapy (RA: DAS28 4.34± 1,39 and SpA: BASDAI 5.51 ±1.61). All patients had a stable clinical activity along both time periods [RA (DAS28): 2.51±0.85 in 1st P vs 2.31±0.52 in 2nd P, p=0.061; SpA (BASDAI): 1.90±1.11 in 1stP vs 1.88±1.12 in 2nd P, p=0.907]. In the 2nd P the drug administration interval was higher for all three TNF inhibitors (Ifx: 8.52 ±1.43 weeks 1st P vs 9.7±1.44 weeks 2nd P p<0,001; Ada: 2.19 ± 0,58 weeks 1stP vs 2.95±1.58 weeks 2ndP p=0,007; Etn: 1.09±0,27 weeks 1stP vs 1.61 ±0,91 weeks 2ndP p=0.004). The weekly mean dose received per patient from each anti-TNF in the 2nd P was significantly lower than in the 1st P [Ifx (mg/kg/week): 0.51±0.14 at 1stP vs 0.42±0.12 at 2ndP, p<0.001; Ada (mg/week): 19.19±3.72 at 1stP vs 15.52±4.81 at 2ndP, p<0,001; Etn (mg/week) 42.09±13.25 at 1stP vs 35.04±13.37 at 2ndP, p=0.009]. The monthly amount of spared drug, converted in economic savings was 91.62 € per patient for Ifx (70kg of mean weight), 324 € per patient for Ada, 257 € per patient for Etn. The yearly amount of money saved in this group of 88 patients was 233.186 €. Conclusions The therapeutic dose monitoring based on serum trough drug levels, used in rheumatic patients, has shown to be a useful tool to support therapeutic clinical practice in addition to enabling cost effective savings of biopharmaceuticals References De Vries, Ann Rheum Dis 2009; Bartelds, Ann Rheum Dis 2007, Pascual-Salcedo, Rheumatology 2011. Acknowledgements Unrestricted Educational Grant by Pfizer Disclosure of Interest None Declared

Background The use of anti-TNF therapy, as Etanercept (Etn), has proven to be effective in patients with rheumatoid arthritis (RA). However, a significant proportion of patients develop clinical inefficacy or adverse effects, resulting in treatment interruption. Recently, it has been reported that Etn serum trough levels are correlated with clinical activity in RA patients, so that, patients with low Eta serum trough levels had a worse clinical response. Objectives To evaluate whether Etn serum trough levels correlate with clinical activity and treatment discontinuation in RA patients treated with Etn. Methods In this ambispective study, 44 RA patients treated with Etn were included. Clinical activity was assessed using the Disease Activity Score 28 (DAS28), clinical improvement by the delta-DAS28 and response to treatment using EULAR criteria at baseline and at 1st, 2nd and > 3rd years of treatment. In every patient Etn was administered subcutaneously at a dose of 50 mg/week and Eta levels were assessed every 6 months by capture ELISA. Blood samples were obtained up to 24 hours before administration of Etn. Statistical analysis was performed using SPSS 11.0. For the statistical study, serum trough levels were divided into low levels (LL<1000 ng/ml) and high (HL ≥ 1000 ng/ml). Results Of the 44 RA patients treated with Etn, 34 (77.2%) were women. The mean age was 60.2 ± 12.7 years and the mean disease duration was 16.1 ± 9.9 years. The baseline clinical activity (DAS28) was 4.9 ± 1.0 and there were no differences between patients who later developed Etn LL and HL (5.3 ± 1.5 with LL vs 4.9 ± 0.8 with HL, p = 0.287). Clinical activity (DAS28) was higher in patients with Etn LL at all studied time points (4.3 ± 0.9 with LL vs 2.6 ± 0.6 with HL at 1st year, p = 0.003; 4.3 ± 1.2 with LL vs 2.5 ± 0.7 with HL at 2nd year, p = 0.002; 4.5 ± 0.9 with LL vs 2.7 ± 0.6 with HL at > 3rd year; p = 0.000). Clinical improvement (delta-DAS28) was lower in patients with Etn LL throughout the study (-0.7 ± 1.1 with LL vs 2.1 ± 0.9 with HL at 1 year, p = 0.006; 0.8 ± 2.1 with LL vs 2.4 ± 0.9 with HL at 2nd year, p = 0.036; 0.86 ± 1.4 with LL vs 2.2 ± 0.9 with HL at > 3rd year, p = 0.003). The drop-out of biological therapy during the study was more frequent in patients with Etn LL [7/13 (53.8%) vs 1/31 (3.2%) with HL, p = 0.000]. Etn serum trough levels (Mdn, P25-P75) were higher in patients with EULAR good (GR) and moderate response (MR) than in non-responders (NR) patients [2287.0; 1488.7-2625.0 classified as GR vs 821.5; 500.0-1143.0 with MR vs. 16.0; 16.0-16.0 with NR at 1st year, p = 0.026; 1379.2; 1190.5-3347.0 with GR vs 2225.0; 2225.0-2225.0 with MR vs 352.0; 0.0-973.0 with NR at 2nd year, p=0.154; 2300.0; 1812.0-3312.0 with GR vs 1274.0; 700.5-2605.0 with MR vs 37.0; 15.0-1200.0 with NR at >3rd year, p=0.007]. Conclusions The presence of low Etn serum trough levels correlates with a poor clinical response to Etn and a higher frequency of treatment discontinuation. However, patients classified as responders have higher Etn serum trough levels and a low frequency of treatment discontinuation. Disclosure of Interest None Declared

Background Spondyloarthropathies (SpA) include a heterogeneous group of rheumatic diseases that mainly affect the axial skeleton and entheses. Despite the anti-TNF therapy has been demonstrated effective in SpA patients, about 30% of them develop primary or secondary inefficacy. In rheumatoid arthritis (RA) has been shown that the development of antibodies (Ab) against the first anti-TNF determines the clinical response to a second anti-TNF. Objectives To assess if the effectiveness of second anti-TNF therapy is associated with the development of Ab against the first anti-TNF in SpA patients switching to a second anti-TNF. Methods We studied 33 SpA patients treated with a second anti-TNF after having an inadequate response to the first anti-TNF therapy. The diagnoses were: 23 (69.7%) Ankylosing Spondylitis (AS), 6 (18.2%) undifferentiated SpA, 2 (6.1%) psoriatic SpA, 1 (3%) SpA associated with inflammatory bowel disease and 1 (3%) reactive SpA. Clinical activity was measured by ASDAS-CRP at baseline to the 1st and 2nd anti-TNF treatment and after 6 months of the switching. Clinical improvement was assessed by the Delta-ASDAS (clinically important improvement ≥1.1). Drug through levels and anti-drug Ab were measured by ELISA at the end of the fist anti-TNF treatment. Statistical analysis was performed using SPSS 11.0. Results Our cohort included 33 patients, 18 (54.5%) male, mean age of 50.09±10 years and 21 (63.6%) HLA B27 positive. All of them were initially treated with an anti-TNF: 14 (42.4%) with infliximab (Ifx), 3 (9.1%) with adalimumab (Ada), 16 (48.5%) with etanercept (Eta). Nine out of 33 (27.3%) developed anti-drug Ab [8 anti-Ifx Ab (ATI) and 1 anti-Ada Ab (AAA)]. All patients switched to a 2nd anti-TNF, due to inefficacy: 7 (21.2%) to Ifx, 16 (48.5%) to Ada, 5 (15.2%) to Eta, 5 (15.2%) to golimumab. Clinical activity (ASDAS-CRP) was not different at baseline between first and second anti-TNF therapy (3.45±0.99 vs 3.22±0.96, p=0.24). No differences were observed between patients who developed or not anti-drug Ab at baseline to the first (3.34±0.87 with Ab vs 3.50±1.04 without Ab, p=0.93) and second (2.99±0.95 with Ab vs 3.31±0.96 without Ab, p=0.37) anti-TNF treatment. After 6 months of switching to a second anti-TNF, patients who had developed anti-drug Ab had lower clinical activity (ASDAS-CRP) than patients without anti-drug Ab (1.76±0.98 with Ab vs 2.79±1.10 without Ab, p=0.021). Moreover clinical improvement was higher in patients with anti-drug Ab (1.23±1.22 with Ab vs 0.52±1.08 without Ab, p=0.063). Most patients who had clinically important improvement had anti-drug Ab before switching [60% (6/10) vs 40% (4/10), p=0.010]. Conclusions Same as in RA, in SpA, the failure to a first anti-TNF therapy associated with the development of anti-drug Ab predicts a better clinical response to a second anti-TNF. The study of the immunogenicity in the biological treatment failure helps predict the response to a second biological treatment in SpA. Disclosure of Interest None Declared

Studies on three continents have reported associations between various measures of airborne particles and daily deaths. Sulfur dioxide has also been associated with daily deaths, particularly in Europe. Questions remain about the shape of those associations, particularly whether there are thresholds at low levels. We examined the association of daily concentrations of black smoke and SO2with daily deaths in eight Spanish cities (Barcelona, Bilbao, Castellón, Gijón, Oviedo, Valencia, Vitoria, and Zaragoza) with different climates and different environmental and social characteristics. We used nonparametric smoothing to estimate the shape of the concentration-response curve in each city and combined those results using a metasmoothing technique developed by Schwartz and Zanobetti. We extended their method to incorporate random variance components. Black smoke had a nearly linear association with daily deaths, with no evidence of a threshold. A 10 μ g/ m3increase in black smoke was associated with a 0.88% increase in daily deaths (95% confidence interval, 0.56%-1.20%). SO2had a less plausible association: Daily deaths increased at very low concentrations, but leveled off and then decreased at higher concentrations. These findings held in both one- and two-pollutant models and held whether we optimized our weather and seasonal model in each city or used the same smoothing parameters in each city. We conclude that the association with particle levels is more convincing than for SO2, and without a threshold. Linear models provide an adequate estimation of the effect of particulate air pollution on mortality at low to moderate concentrations.