Comparing pain intensity rating scales in acute postoperative pain: boundary values and category disagreements.

Pain intensity assessment scales are important in evaluating postoperative pain and guiding management. Different scales can be used for patients to self-report their pain, but research determining cut points between mild, moderate and severe pain has been limited to studies with < 1500 patients. We examined 13,017 simultaneous acute postoperative pain ratings from 913 patients taken at rest and on activity, between 4 h and 48 h following surgery using both a verbal rating scale (no, mild, moderate or severe pain) and 0-100 mm visual analogue scale. We determined the best cut points on the visual analogue scale between mild and moderate pain as 35 mm, and moderate and severe pain as 80 mm. These remained consistent for pain at rest and on activity, and over time. We also explored the presence of category disagreements, defined as patients verbally describing no or mild pain scored above the mild/moderate cut point on the visual analogue scale, and patients verbally describing moderate or severe pain scored below the mild/moderate cut point on the visual analogue scale. Using 30 and 60 mm cut points, 1533 observations (12%) showed a category disagreement and using 35 and 80 mm cut points, 1632 (13%) showed a category disagreement. Around 1 in 8 simultaneous pain scores implausibly disagreed, possibly resulting in incorrect pain reporting. The reasons are not known but low rates of literacy and numeracy may be contributing factors. Understanding these disagreements between pain scales is important for pain research and medical practice.

The global prevalence of primary psychodermatologic disorders: a systematic review.

The management of primary psychodermatologic disorders (PPDs) (i.e. psychiatric disorders with dermatologic presentation) is challenging. The scarceness of reported prevalence hinders the development of coordinated interventions to improve healthcare delivery. This review aimed to explore the global prevalence of PPDs. The review was conducted according to the Preferred Reporting Items for Systematic Review and Meta-Analyses statement. Of the 4632 identified publications, 59 were included. Five PPDs were investigated from the included studies: delusional parasitosis (n = 9), skin picking disorder (n = 26), trichotillomania (n = 26), tanning dependence (n = 5) and repetitive nail biting (n = 6). Delusional parasitosis was rare in the general population (prevalence ranging from 0.0002% to 0.03%), with higher rates in psychiatric settings (outpatient = 0.5%; inpatient = 0.1%). Other pathologic or subclinical forms of PPDs had a minimum prevalence of 0.3% (median = 7.0%; mean = 17.0%). The distribution of the prevalence rates was highly skewed, with large differences based on the study setting (e.g. dermatologic settings, psychiatric settings, and general population). The most common condition was pathologic skin picking (prevalence, 1.2%-11.2%) in the general population. Its rates were higher in the psychiatric settings (obsessive-compulsive disorder, 38.5%; Tourette syndrome, 13.0%; body dysmorphic disorder, 26.8%-64.7%). The prevalence of trichotillomania in the general population ranged from 0.6% to 2.9%, while that of pathologic tanning and nail biting could not be ascertained as the studies were mainly in students (range; 12.0%-39.3% and 3.0%-10.1%, respectively). In conclusion, PPDs are common, especially in the dermatologic and psychiatric settings. Further population-based studies are needed to determine more accurate prevalence rates.

Factors predicting persistence of biologic drugs in psoriasis: a systematic review and meta-analysis.

BACKGROUND: Long-term therapy for psoriasis is impaired by gradual loss of effectiveness and treatment discontinuation. Identifying factors that affect biologic drug survival may help in treatment optimization. OBJECTIVES: To identify factors that predicted biologic drug persistence or discontinuation in a real-life setting. METHODS: We identified studies of biologic persistence in psoriasis through a comprehensive, systematic literature search using predefined search criteria. Studies were screened by title and abstract then further by full-text review. Hazard ratio (HR) data were extracted for all available predictive factors (HRs > 1 denoted biologic discontinuation, and HRs < 1 denoted biologic persistence). A meta-analysis of HRs (random-effects model) was used to assess any predictive factor included in at least two studies. RESULTS: Sixteen cohort studies were included in the review, with a total of 32 194 patients. A meta-analysis was performed on 13 studies (n = 29 802): nine for female sex (n = 28 090), six for obesity (n = 9311) and six for psoriatic arthritis (n = 24 444). Obesity and female sex predicted treatment discontinuation, with HRs of 1.21 [95% confidence interval (CI) 1.10-1.32, I2 = 0%] and 1.22 (95% CI 1.07-1.38, I2 = 84%), respectively. Concomitant psoriatic arthritis predicted biologic persistence (HR 0.83, 95% CI 0.80-0.86, I2 = 0%). Female sex predicted biologic discontinuation due to side-effects, with a pooled HR of 2.16 (95% CI 1.39-3.35, I2 = 67%). Other reported predictive factors (smoking, metabolic syndrome, biologic naivety, age, Dermatology Life Quality Index, dyslipidaemia, high socioeconomic status and concomitant methotrexate) were insufficiently reported for meta-analysis. CONCLUSIONS: Our meta-analysis demonstrates that female sex and obesity predict biologic discontinuation, and concomitant psoriatic arthritis predicts biologic survival. What's already known about this topic? Ineffectiveness is the main factor that causes drug discontinuation during long-term treatment of psoriasis. It is unclear which factors and comorbidities impact drug persistence. What does this study add? Female sex and obesity predict biologic discontinuation due to ineffectiveness and adverse events. Concomitant psoriatic arthritis is associated with improved drug persistence.

Co-administration of dexamethasone with peripheral nerve block: intravenous vs perineural application: systematic review, meta-analysis, meta-regression and trial-sequential analysis.

BACKGROUND: I.V. and perineural dexamethasone have both been found to prolong loco-regional analgesia compared with controls without dexamethasone. It is unclear whether perineural administration offers advantages when compared with i.v. dexamethasone. METHODS: A systematic literature search was performed to identify randomized controlled double-blind trials that compared i.v. with perineural dexamethasone in patients undergoing surgery. Using the random effects model, risk ratio (for binary variables), weighted mean difference (for continuous variables) and 95% confidence intervals were calculated. We applied trial sequential analysis to assess the risks of type I and II error, meta-regression for the study of the doseresponsive relationship, and the Grading of Recommendations Assessment, Development, and Evaluation system. RESULTS: We identified 10 randomized controlled double-blind trials (783 patients). When using conventional meta-analysis of nine low risk of bias trials, we found a statistically significantly longer duration of analgesia, our primary outcome with perineural dexamethasone (241 min, 95%CI, 87, 394 min). When trial sequential analysis was applied, this result was confirmed. Meta-regression did not show a dose-response relationship. Despite the precision in the results, using the Grading of Recommendations Assessment, Development, and Evaluation system (GRADE), we assessed the quality of the evidence for our primary outcome as low. CONCLUSIONS: There is evidence that perineural dexamethasone prolongs the duration of analgesia compared with i.v. dexamethasone. Using GRADE, this evidence is low quality.

Paravertebral block and persistent postoperative pain after breast surgery: meta-analysis and trial sequential analysis.

We examined whether paravertebral block has an effect on the prevalence of persistent postsurgical pain after breast surgery. Seven randomised, controlled trials (559 patients) which had the outcome assessor blinded were included, comparing patients who received paravertebral blocks after breast surgery with patients who did not. The risk ratio (95% CI) was 0.75 (0.48-1.15) for the incidence of postoperative pain at 3 months (four studies, 317 patients); the risk ratio (95% CI) obtained from three studies including 301 patients reporting on pain after 6 months was 0.57 (0.29-1.72), and the risk ratio (95% CI) for pain after 12 months (three trials, 237 patients) was 0.42 (0.15-1.23). Conventional meta-analysis using the random effects model thus showed no statistically significant risk reduction for persistent postoperative pain at 3 months, 6 months or 12 months. Trial sequential analysis, used to consider the risk of type 1 and type 2 random error, showed that at 3 months, 6 months and 12 months, the number of subjects in the analyses were only 18.3%, 6.8% and 4.2% of the required information sizes at those time points, respectively. Our study is the first to evaluate data on pain 12 months postoperatively. Trial sequential analysis revealed that the current evidence is not sufficient to reach a conclusion. These findings stand in contrast to previous meta-analyses with fewer studies that had concluded that paravertebral block effectively reduces chronic pain.

[Percentile Values for the Anthropometric Dimensions of Triplet Neonates - Analysis of German Perinatal Survey Data of 2007-2011 from all States of Germany]. / Perzentilwerte für die Körpermaße neugeborener Drillinge - Ergebnisse der deutschen Perinatalerhebung der Jahre 2007-2011 unter Beteiligung aller Bundesländer.

AIM: We aimed to develop national reference values for birth weight, length, head circumference, and weight for length for newborn triplets based on data from the German perinatal survey of 2007-2011. MATERIAL AND METHODS: Perinatal survey data of 3,690 newborn triplets from all the states of Germany were kindly provided to us by the AQUA Institute in Göttingen, Germany. Data of 3,567 newborn triplets were included in the analyses. Sex-specific percentile values were calculated using cumulative frequencies. Percentile values at birth were computed for the 3rd, 10th, 25th, 50th, 75th, 90th, and 97th percentiles for 21-36 completed weeks of gestation. RESULTS AND CONCLUSIONS: We present the first German reference values (tables and curves) for the anthropometric dimensions of triplet neonates and compare selected birth weight and length percentiles of triplets (after 32 and 34 completed weeks of gestation) to those of singletons and twins. The differences in the 50th birth weight percentiles between singletons and triplets after 32 completed weeks of gestation were 180 g for girls and 210 g for boys; after 34 weeks of gestation the differences were 320 and 325 g, respectively. The differences between twins and triplets after 32 weeks of gestation were 100 g for girls and 120 g for boys; after 34 weeks of gestation they were 130 and 135 g, respectively. The data presented here enable the classification of newborn triplets according to somatic parameters making reference to German perinatal data.

Intravenous ketamine during spinal and general anaesthesia for caesarean section: systematic review and meta-analysis.

BACKGROUND: Intravenous ketamine has been used during general and regional anaesthesia for caesarean section. No systematic review and meta-analysis on the desired effects and adverse effects of ketamine administration during caesarean section have yet been performed. METHODS: After a systematic literature search a meta-analysis was conducted with the random effects model. Weighted mean difference (WMD) or risk ratio and 95% confidence intervals (CIs) were computed. RESULTS: Twelve randomised controlled double-blind trials comprising 953 patients were included: seven studies reported on spinal anaesthesia and five on general anaesthesia. Significant differences in the aforementioned outcome variables were found only in the spinal anaesthesia studies. In the spinal anaesthesia studies the time to the first analgesic request was significantly longer in ketamine-treated women, the WMD was 49.36 min (95% CI 43.31-55.41); visual analogue scale pain scores at rest 2 h after surgery were significantly lower. No differences were observed for maternal nausea, vomiting, pruritus, and psychomimetic effects. Only few data were found for neonatal outcomes. CONCLUSIONS: We conclude that ketamine enhances post-operative analgesia after caesarean section under spinal anaesthesia. There is a paucity of data for several maternal adverse effects as well as for neonatal well-being. Further studies are needed for general anaesthesia.

Prophylactic tranexamic acid in parturients at low risk for post-partum haemorrhage: systematic review and meta-analysis.

Tranexamic acid is effective in reducing blood loss during various types of surgery and after trauma. No compelling evidence has yet been presented for post-partum haemorrhage. A systematic literature search of relevant databases was performed to identify trials that assessed blood loss and transfusion incidence after tranexamic acid administration for post-partum haemorrhage. The random effects model was used for meta-analysis. Risk ratios (RRs) and weighted mean differences (WMDs) were calculated with 95% confidence intervals (CIs). Seven trials with a low risk of bias comparing tranexamic acid vs. placebo with a total of 1760 parturients were included in our systematic review and meta-analysis. Blood loss was significantly lower after tranexamic acid use (WMD -140.29 ml, 95% CI -189.64 to -90.93 ml; P<0.00001). Tranexamic acid reduced the risk for blood transfusions (RR 0.34, 95% CI 0.20-0.60, P=0.0001). The incidence of transfusions in the placebo group varied between 1.4% and 33%. When omitting the two trials with the highest incidence of transfusions, the RR was no longer significant. Additional uterotonics were necessary in the placebo groups; gastrointestinal adverse events were more common after tranexamic acid use. Only four cases of thrombosis were found, two each in the tranexamic acid and control groups. Tranexamic acid effectively reduced post-partum blood loss; the effect on the incidence of blood transfusions requires further studies. Only few trials observed adverse events including thromboembolic complications and seizures.

Prophylactic phenylephrine for caesarean section under spinal anaesthesia: systematic review and meta-analysis.

We conducted a systematic review to determine the harm and benefit associated with prophylactic phenylephrine for caesarean section under spinal anaesthesia. We included 21 randomised controlled trials with 1504 women. The relative risk (95% CI) of hypotension with phenylephrine infusion ­ as defined by authors ­ before delivery was 0.36 (0.18­0.73) vs placebo, p = 0.004; 0.58 (0.39­0.88) vs an ephedrine infusion, p = 0.009; and 0.73 (0.55­0.96) when added to an ephedrine infusion, p = 0.02. After delivery, the relative risks of hypotension and nausea and vomiting with phenylephrine compared with placebo were 0.37 (0.19­0.71), p = 0.003, and 0.39 (0.17­0.91), p = 0.03, respectively. There was no evidence that hypertension, bradycardia or neonatal endpoints were affected. Phenylephrine reduced the risk for hypotension and nausea and vomiting after spinal doses of bupivacaine generally exceeding 8 mg, but there was no evidence that it reduced other maternal or neonatal morbidities.

Meta-analysis of the success of block following combined spinal-epidural vs epidural analgesia during labour.

Observational studies suggest that combined spinal-epidural analgesia (CSE) is associated with more reliable positioning, lower epidural catheter replacement rates, and a lower incidence of unilateral block compared with epidural analgesia. However, evidence from high-quality trials still needs to be assessed systematically. We performed a systematic review that included 10 randomised controlled trials comparing CSE and epidural analgesia in 1722 labouring women in labour. The relative risk of unilateral block was significantly reduced after CSE vs epidural analgesia (0.48, 95% CI 0.24-0.97), but significant between-study heterogeneity was present (I(2) = 69%, p = 0.01). No differences were found for rates of epidural catheter replacement, epidural top-up, and epidural vein cannulation. On the basis of current best evidence, a consistent benefit of CSE over epidural analgesia cannot be demonstrated for the outcomes assessed in our review. A large randomised controlled trial with adequate power is required.

[New percentile values for the anthropometric dimensions of singleton neonates: analysis of perinatal survey data of 2007-2011 from all 16 states of Germany]. / Neue Perzentilwerte für die Körpermaße neugeborener Einlinge: Ergebnisse der deutschen Perinatalerhebung der Jahre 2007-2011 unter Beteiligung aller 16 Bundesländer.

AIM: The aim of this study was to derive percentile values for birth weight, length, head circumference, and weight for length for singleton neonates based on the German perinatal survey of 2007-2011 (using data from all 16 states of Germany). We also compared these new percentile values with the percentile values of 1995-2000 that so far have been considered standard values. MATERIAL AND METHODS: Data of 3 187 920 singleton neonates from the German perinatal survey of the years 2007-2011 were kindly provided to us by the AQUA Institute in Göttingen, Germany. Sex specific percentile values were calculated using cumulative frequencies. Percentile values at birth were computed for the 3(rd), 10(th), 25(th), 50(th), 75(th), 90(th), and 97(th) percentiles for 21-43 completed weeks of gestation. Percentile curves and tabulated values for the years 2007-2011 were compared with the published values of 1995-2000. RESULTS AND DISCUSSION: Overall the new percentile curves closely resemble the previous ones. Minimal differences can be found for the 10(th) percentile and generally for early weeks of gestation. Values for the 10(th) percentile in the 2007-2011 dataset are somewhat higher than values of 1995-2000 for birth weight, length, and weight for length. CONCLUSIONS: We recommend the use of these new percentile values instead of the old ones.

[New percentile values for the anthropometric dimensions of twin neonates: analysis of perinatal survey data of 2007-2011 from all 16 states of Germany]. / Neue Perzentilwerte für die Körpermaße neugeborener Zwillinge: Ergebnisse der deutschen Perinatalerhebung der Jahre 2007-2011 unter Beteiligung aller 16 Bundesländer.

AIM: The aim of this study was to develop new national standards for birth weight, length, head circumference, and weight for length for newborn twins based on the German perinatal survey of 2007-2011. We also assessed trends in anthropometric measurements by comparing these new percentile values with the percentile values of 1990-1994. MATERIAL AND METHODS: Perinatal survey data of 110,313 newborn twins from all the states of Germany collected in the years 2007-2011 were kindly provided by the AQUA Institute in Göttingen, Germany. Sex specific percentile values were calculated using cumulative frequencies. Percentile values at birth were computed for the 3rd, 10th, 25th, 50th, 75th, 90th, and 97th percentiles for 21-40 completed weeks of gestation. Percentile curves and tabulated values for the years 2007-2011 were compared with the published values of 1990-1994. RESULTS AND DISCUSSION: The new percentile curves (2007-2011) closely resemble the previous ones (1990-1994). Small differences can nonetheless be found. For example, for birth weight the new values for the 10th percentile are a little higher. CONCLUSIONS: We recommend using the new percentile values instead of the old ones.

Maternal and foetal effects of remifentanil for general anaesthesia in parturients undergoing caesarean section: a systematic review and meta-analysis.

BACKGROUND: Remifentanil has been suggested for the induction of general anaesthesia for caesarean section. We aimed to define remifentanil effects on maternal stress response as well as neonatal effects. METHODS: Relevant articles were retrieved by a systematic literature search. Randomized, controlled trials comparing remifentanil use before delivery with placebo were selected. Maternal outcome parameters were blood pressure and heart rate; neonatal effects included the need for mask ventilation and intubation, base excess, pH values, Apgar < 7 at 1 and 5 min. The random effects model was used for meta-analysis; risk ratio or weighted mean difference (WMD) and 95% confidence interval (95% CI) were calculated. RESULTS: Five articles including 186 patients were identified. Highest and lowest systolic blood pressure were significantly lower in the remifentanil group (WMD: -29.98, -50.90 to -9.07 mmHg, 95% CI; P = 0.005; and WMD: -12.46, -18.21 to -6.71 mmHg, 95% CI; P < 0.0001), the lowest heart rate was significantly lower after remifentanil treatment (WMD: -8.22, -11.67 to -4.78, 95% CI; P < 0.00001). Base excess was significantly higher in infants of remifentanil-treated mothers (WMD: 1.15, -0.27 to 2.03, 95% CI; P = 0.01); pH was also higher in the remifentanil group, but significance was missed (P = 0.07). No differences were observed for Apgar values or the need of airway assist. CONCLUSION: Remifentanil was found to attenuate the maternal circulatory response to intubation and surgery. Higher base excess and pH suggest a beneficial effect on the neonatal acid-base status. A trial with adequate power is warranted that addresses neonatal side-effects of remifentanil.

Effects of smoking specified as cigarettes per day and maternal body mass index on hypertensive disorders of pregnancy.

BACKGROUND: We have previously described the prevalence in pregnancy of hypertension, proteinuria, oedema and preeclampsia/eclampsia according to maternal body mass index (BMI) and smoking status. We found that these disorders were less frequent among smoking women. To investigate whether this relationship is causal or a chance finding, we here present an analysis according to BMI and smoking specified according to the number of cigarettes consumed per day. MATERIALS AND METHODS: Data were from the German Perinatal Survey of 1998-2000. We classified women by BMI as underweight (BMI<18.5 kg/m2), normal weight (BMI 18.5-24.99 kg/m2), overweight (25.0-29.99 kg/m2), or obese (BMI≥30 kg/m2). Smoking was categorised as being a non-smoker or smoking 1-7, 8-14 or ≥ 15 cigarettes per day. Datasets from 433 669 singleton pregnancies with information on maternal BMI and smoking were included in the analysis. RESULTS: In all BMI categories hypertension, moderate to severe oedema, and preeclampsia/eclampsia became less prevalent with increasing maternal cigarette consumption. CONCLUSIONS: Dose-dependence was not convincing for proteinuria.Dose-dependence in the relationship between smoking and hypertensive disorders of pregnancy argues against a chance finding and for a causal relationship.

Analysis of the German perinatal survey of the years 2007-2011 and comparison with data from 1995-1997: neonatal characteristics and duration of pregnancy.

BACKGROUND AND AIM: We have previously analysed neonatal characteristics and duration of pregnancy in Germany based on data from the German Perinatal Survey of 1995-1997. Here we describe neonatal characteristics and duration of pregnancy based on the German Perinatal Survey of 2007-2011. MATERIAL AND METHODS: We had been provided with data from the German Perinatal Survey of 1995-1997 by the chambers of physicians of all the states of Germany except Baden-Württemberg (1 815 318 singleton neonates). We were also provided with access to the perinatal survey data of 2007-2011 by the AQUA Institute in Göttingen, Germany (3 187 920 singleton neonates). We investigated regional differences within Germany and also compared the 2 periods of time. We used the computer programme SPSS for data analysis and performed plausibility checks on the survey data. RESULTS: Comparing the states of Germany, we found that birth weight was largest for neonates born in Schleswig-Holstein (3 407 g) and Mecklenburg-Western Pomerania (3 392 g); the lowest mean birth weight was observed in the Saarland (3 283 g). Preterm birth rate varied between 6.3% (Saxony) and 8.1% (Bremen, Saarland). Comparing 1995-1997 vs. 2007-2011, deliveries after 37 and 38 weeks of gestation were more common and deliveries after 39 and more weeks of gestation were less common in the later period of time. CONCLUSIONS: Regional differences in the anthropometric characteristics of neonates exist between the states of Germany. The proportion of deliveries after 39 and more weeks of gestation has decreased.

On weight gain during pregnancy: relationships between weight gain during pregnancy, duration of pregnancy and the somatic classification of neonates.

BACKGROUND: Weight gain during pregnancy is an important parameter that is related to a number of perinatal outcomes. We aimed to analyse the relationships between weight gain during pregnancy, duration of pregnancy, and the somatic classification of the neonates as small, appropriate, or large for gestational age (SGA, AGA, LGA). MATERIAL AND METHODS: Data were from the German perinatal survey of 1995-2000 (more than 2.2 million singleton pregnancies). We classified all neonates with a birth weight below the 10th population percentile as SGA, those with a birth weight above the 90th percentile as LGA, and all others were AGA. Duration of pregnancy (categorised as ≤ 36, 37-41, or ≥ 42 completed weeks of gestation) and the percentages of SGA, AGA, and LGA neonates were analysed according to maternal weight gain in 1-kg-steps. RESULTS: Small weight gain was associated with higher rates of preterm birth, i.e. birth after ≤ 36 completed weeks of gestation (preterm birth rates >10% for women who gained <9 kg). SGA rates were greater for low weight gain values and LGA rates were greater for high weight gain values. For weight gains <12 kg, SGA rates were always >10%. For weight gains >14 kg LGA rates were always >10% reaching LGA rates >25% for weight gains in the range 33-35 kg. CONCLUSIONS: Weight gain during pregnancy may be of use as a predictor of perinatal outcomes such as the somatic classification of neonates. Further analyses taking account of factors influencing the weight gain during pregnancy are warranted.

On the variability in preterm birth rate, birth weight, and somatic classification among neonates of mothers with the same body mass index.

BACKGROUND: Maternal height and weight are important determinants of perinatal outcomes.Height and weight can be combined in the measure of body mass index (BMI). We aimed to investigate the utility of maternal BMI as a predictor of perinatal outcomes. MATERIALS AND METHODS: Based on data collected between 1995 and 2000 as part of the German perinatal survey, we examined singleton pregnancies of women with BMIs of 18, 24, or 30. We compared preterm birth rate, birth weight, and the somatic classification of neonates as small,appropriate, or large for gestational age (SGA,AGA, LGA) for women with heights of 150 cm and 180 cm for each BMI. RESULTS: For women with a BMI of 18 (24; 30)and a height of 150 cm, the preterm birth rate was 13.9 % (9.1 %; 12.5 %); for women with the same BMI and a height of 180 cm the preterm birth rate was 12.1 % (6.1 %; 4.4 %). Birth weight for women with a BMI of 18 (24; 30) and a height of 150 cm was 2 889 g (3 170 g; 3 147 g); for women with the same BMI and a height of 180 cm it was 3 314 g (3 629 g; 3 753 g). The LGA rate for women with a BMI of 18 (24; 30) and a height of 150 cm was 2.1 % (5.2 %; 5.2 %); for women with the same BMI and a height of 180 cm it was 7.7 %(20.5 %; 27.7 %). CONCLUSIONS: There is considerable variability in perinatal outcomes between women with the same BMI but different heights. This limits the utility of BMI as a predictor of perinatal outcomes.

The combined effect of maternal body mass index and smoking status on perinatal outcomes - an analysis of the german perinatal survey.

BACKGROUND: Maternal body mass index (BMI) outside the normal range and smoking are both associated with adverse perinatal outcomes, but their interaction needs further investigation. AIM: The aim of this study was to analyse the combined effects of smoking and BMI on birth weight, preterm birth rate, the somatic development of neonates, and complications of pregnancy. MATERIAL AND METHODS: Data from 508 926 singleton pregnancies from the German Perinatal Survey of 1998-2000 were analysed according to maternal BMI and smoking. RESULTS: Preterm birth rates were higher for non-smoking underweight (8.3%) and obese women (6.7%) than for normal weight (6.0%) or overweight women (5.6%); rates were higher in smokers than in non-smokers for every BMI category. The mean birth weight increased with increasing BMI and was decreased by smoking; it was 2,964 g in underweight smokers and 3,556 g in obese non-smokers. Small for gestational age (SGA) rates were least in obese women and highest in underweight women; large for gestational age (LGA) rates varied in the opposite direction. In smokers SGA rates were higher than in non-smokers for every BMI category and LGA rates were always lower. Hypertension, proteinuria, oedema, and pre-eclampsia/eclampsia were more common as BMI increased but were always lower in smokers. Pre-eclampsia/eclampsia occurred in 0.7% of underweight smokers but in 9.6% of obese non-smokers. CONCLUSIONS: Smoking and low maternal BMI in combination can cause high rates of preterm birth and SGA neonates as well as low mean birth weight. Although smoking offers some apparent benefit regarding LGA rates and pre-eclampsia this should not distract from its overall adverse influence.