Infradiaphragmatic versus supradiaphragmatic Hodgkin lymphoma: a retrospective review of 1,114 patients.

Infradiaphragmatic Hodgkin lymphoma (IDH) accounts for 4-13% of cases of stage I-II Hodgkin lymphoma (HD). It has been associated with distinct pre-treatment characteristics and outcomes when compared with supradiaphragmatic HD (SDH). The comparison of IDH vs SDH can only be made in early and intermediate stages (I-II), such a comparison is not possible for advanced stages (III-IV). This study retrospectively compared two groups of 1013 patients with stage I-II SDH and 101 patients with IDH (10%). These two sub-groups of patients were treated in 1988-1993 in 2 prospective randomized clinical trials in Germany for early and intermediate stages of Hodgkin lymphoma. IDH-patients were older (median 39 vs 31 years; p < 0.001), predominantly male (73% vs 52%; p < 0.001) and more often had involvement of 3 lymph node areas (LNA) (80% vs 55%; p < 0.001). Histology in IDH was more likely to be mixed cellularity (46.5% vs 23.6%, p < 0.001) or lymphocyte predominant (20 vs 10%, p = 0.003) and less likely nodular sclerosis (25% vs 63%, p < 0.001). In early-stage unfavorable disease, IDH was associated with a higher treatment failure rate (unadjusted hazard ratio 2, 95% CI, 1.3-3.4; p = 0.003). After controlling for age, sex, stage, histology, B-symptoms and involvement of 3 LNA, the adjusted hazard ratio was 1.25 (95% CI, 0.65-2.4; p = 0.51) so that IDH was no longer associated with a statistically significant treatment failure rate. Poorer outcomes with IDH as compared to SDH are attributable to its association with known adverse prognostic risk factors, but IDH, in itself, is not an independent adverse prognostic factor for treatment failure or survival.

The impact of diabetes and obesity on liver histology in patients with hepatitis C.

AIM: An association between diabetes mellitus and HCV has been recognized previously. No study has examined whether there is an independent association between the degree of hepatic fibrosis and the incidence of diabetes in HCV patients when controlling for other risk factors. METHODS: We reviewed the charts of 264 consecutive patients with chronic HCV infection at a referral liver centre from January 1991 to December 1999. Demographic background, medical history, laboratory and liver biopsy results were retrieved. RESULTS: The prevalence of diabetes was 16.3%. Gender, intravenous drug use, steatosis scores, aminotransferase levels and iron studies were similarly distributed in patients with and without diabetes (all p > 0.05). In contrast, mean age was greater in the diabetic group (49.8 vs. 44.3, p = 0.003). The prevalence of diabetes was substantially higher in African-Americans (p = 0.001) and those with BMI > 30 (p = 0.015). Although the fibrosis score was higher in diabetics (rho = 0.14, p = 0.03), that association did not remain significant when controlling for diabetes risk factors (p > 0.3). The degree of steatosis and fibrosis both tended to increase with increasing BMI (rho = 0.47, p < 0.001 and rho = 0.13, p = 0.03, respectively). Even after controlling for diabetes, age, gender, race, and current alcohol use, those associations remained (both p < 0.001). CONCLUSIONS: The prevalence of diabetes in our group of HCV patients was high, consistent with other studies. Diabetes is not an independent predictor of degree of fibrosis. Body mass index is an independent predictor of both fibrosis and steatosis in HCV patients.

How gastroenterologists inform patients of results after lower endoscopy.

OBJECTIVE: Little is known about how gastroenterologists communicate endoscopic findings and biopsy results to their patients. We sought to determine the factors that may influence this behavior. METHODS: A survey questionnaire was developed and mailed to the 80 members of the Delaware Valley Society for GI Endoscopy. Information was obtained on the demographic characteristics and responses to six case vignettes prepared to examine communication patterns. We determined possible influences of conscious sedation and the benignity or severity of findings on communication practices. RESULTS: Sixty-one surveys (76%) were completed and analyzed. Endoscopists immediately inform patients of normal results. For abnormal results, 92% would immediately inform nonsedated patients versus 79% that would inform sedated patients (p < 0.008). Analysis of responses to the case vignettes indicated that 82% of endoscopists would immediately reassure the patient about a benign appearing (< 1 cm) polyp, but only 70% would do so for a polyp > 2 cm (p < 0.01). In contrast, when presented with a frank malignancy, 94% would inform the patient. Eighty-four percent of endoscopists would telephone results of a benign pathology report, but only 34% would telephone report a dysplastic lesion (p < 0.001). There was no correlation between the response rate and various demographic parameters such as physician age, type of, or length of time in practice. CONCLUSIONS: Gastroenterologists usually report normal findings immediately, but are less likely to do so after use of sedation or encountering abnormal findings. Most of those surveyed would use the telephone to communicate abnormal findings.

Intravenous magnesium is ineffective in adult asthma, a randomized trial.

Intravenous magnesium sulphate (MgSO4) has been tried in the emergency department treatment of asthma since the mid-1980s, but published reports vary as to its efficacy. The literature suggests that it may be effective in the more severely ill asthmatic. We evaluated i.v. MgSO4 in adult asthmatics having a moderate to severe exacerbation. The study was performed in a convenience sample of adult asthmatics between the ages of 18 and 55 presenting to the emergency department with a peak expiratory flow (PEF) of < 100 l/min or < 25% of predicted flow. Patients received either 2.0 grams of MgSO4 or placebo in a randomized, double-blind fashion. All patients received inhaled bronchodilators and i.v. steroids. Outcome variables were: improvement in PEF, subjective respiratory distress as measured by the Borg dyspnoea scale (BDS) and hospital admission. The first visits of 42 patients presenting with acute asthma exacerbations were evaluated, 18 receiving MgSO4 and 24 receiving placebo. The t = 60 peak flow in the MgSO4 group was 174 l/min versus 212 l/min in placebo, p = 0.04. Controlling for age, heart rate, initial PEF and initial BDS in ordinal logistic regression, the t = 60 Borg scale of subjective dyspnoea had an odds ratio of 1.54 in favour of more dyspnoea in MgSO4 (95% C.I., 0.36-6.67; p = 0.56). Five of 18 patients (28%) receiving MgSO4 were admitted compared with 5 of 24 (21%) receiving placebo (p = 0.72). In moderately severe adult asthmatics, 2.0 grams of MgSO4 i.v. resulted in less improvement in peak expiratory flow compared with placebo. MgSO4 did not appear to decrease subjective dyspnoea or the hospital admission rate. This evidence does not support the use of MgSO4 in the treatment of acute asthma.

Tacrolimus conversion improves hyperlipidemic states in stable liver transplant recipients.

With improvements in surgical technique and the advent of new and more effective immunosuppressive agents, survival rates in liver transplant recipients have dramatically improved. However, hyperlipidemia frequently develops in patients administered cyclosporine-based immunosuppression long-term, although it appears to occur less often with newer, tacrolimus-based regimens. We sought to determine whether an isolated change in the baseline immunosuppressive regimen (cyclosporine to tacrolimus) would improve hyperlipidemic states in these patients. Twenty-one long-term stable liver transplant recipients with hyperlipidemia, manifested by elevated cholesterol and/or triglyceride levels, were offered conversion to tacrolimus from cyclosporine A therapy. Lipid profiles were monitored at baseline (while on cyclosporine therapy) and at 1 and 3 months after conversion to tacrolimus therapy. There were no other medication manipulations. After conversion to tacrolimus therapy, mean cholesterol levels decreased from 251 to 202 mg/dL at 1 month (P <.001) and 194 mg/dL at 3 months (P <.001). Similarly, triglyceride levels decreased from 300 to 207 mg/dL by 1 month (P =.011) and 203 mg/dL by 3 months (P <.001). There was also a statistically significant decrease for very low-density lipoprotein levels at 3 months (P =.005) and low-density lipoprotein levels at 1 and 3 months (P =.013 and P =.014, respectively). High-density lipoprotein levels did not significantly change after conversion to tacrolimus therapy. Conversion was not accompanied by adverse side effects, and patients tolerated the change well. In conclusion, simple conversion from cyclosporine to tacrolimus-based immunosuppression therapy is safe and improves posttransplantation hyperlipidemia in a subgroup of liver transplant recipients.

The medical and financial costs associated with termination of a nutrition support nurse.

BACKGROUND: Cost-containment pressures have adversely affected hospital nutrition support team staffing. We determined the effect of termination of a nutrition support nurse responsible for patients receiving total parenteral nutrition (TPN) on quality assurance and financial indicators. METHODS: A retrospective review of all 1,093 patients receiving TPN from fiscal year (FY) 1992 through FY 1998 in a tertiary care community hospital. We documented the changes in care during years when the nutrition support nurse position was staffed, terminated, and restored. Indicators studied included inappropriate TPN, central venous line sepsis, TPN wastage, and estimates of preventable costs. RESULTS: When the nurse was present, 8.6% of TPN patients had a functional gastrointestinal (GI) tract and inappropriately received TPN compared with 12.1% when the nurse was absent, a risk difference of 3.5% points (95% confidence interval [CI], -.06 to 8.3; p = .069). Risk of TPN-associated line sepsis increased from 8.8% of patients when the nurse was present to 13.2% when the nurse was absent, a difference of 4.4% points (95% CI, 0.06 to 9.2; p = .028). In the absence of the nurse, 26.3% of TPN patients had preventable charges vs 17.5% when the nurse was present (p < .0001). Total preventable charges were higher in the years without a nurse (p < .003). Total preventable costs increased by $38,148 to $194,285 (depending on the estimate for sepsis) in the year after termination. Reinstatement of the nurse resulted in a decrease in costs between $34,485 and $156,654. CONCLUSIONS: Adequate staffing of a nutrition support team reduced inappropriate TPN and complications of TPN. Financial savings of the same order of magnitude as the nurse's compensation accompany substantial decreases in patient morbidity.

Hypertonic saline treatment of severe hyperkalemia in nonnephrectomized dogs.

OBJECTIVES: To determine whether a hypertonic saline bolus improves cardiac conduction or plasma potassium levels more than normal saline infusion within 15 minutes of treatment for severe hyperkalemia. Previously with this model, 8.4% sodium chloride (NaCl) and 8.4% sodium bicarbonate (NaHCO(3)) lowered plasma potassium equally effectively. METHODS: This was a crossover study using ten conditioned dogs (14-20 kg) that received, in random order, each of three intravenous (IV) treatments in separate experiments at least one week apart: 1) 2 mmol/kg of 8.4% NaCl over 5 minutes (bolus); 2) 2 mmol/kg of 0.9% NaCl over one hour (infusion); or 3) no treatment (control). Using isoflurane anesthesia and ventilation (pCO(2) = 35-40 torr), 2 mmol/kg/hr of IV potassium chloride (KCl) was infused until conduction delays (both absent p-waves and >/=20% decrease in ventricular rate in

Decalcification and bond failure: A comparison of a glass ionomer and a composite resin bonding system in vivo.

Decalcification during orthodontic treatment is a serious problem. A glass ionomer agent is now available to bond orthodontic brackets as an alternative to composite resins. This prospective study was a clinical trial to determine if a glass ionomer bonding system (Fuji Ortho LC) decreases the incidence of decalcification without increasing the amount of bonding failures. A prospective clinical trial with 16 patients encompassing a total of 298 teeth was conducted. The 149 control teeth were bonded with a light-cured composite resin (Reliance Light Bond) whereas the 149 experimental teeth were bonded with the light-cured glass ionomer agent. Patients were followed for a period of 12 to 14 months. All teeth were evaluated for breakage (bonding failure rate), and all maxillary anterior teeth (96) were evaluated for decalcification on a graded scale. The glass ionomer failure rate was 24.8%, and was higher than the composite resin failure rate of 7.4% (P <.001). There were more glass ionomer bond failures in 12 of 16 patients (P <.001). Enamel decalcification was similar in the 2 bonding systems.

A pilot study of cancer knowledge and screening behaviors of Vietnamese and Cambodian women.

Breast and cervix cancer screening behaviors, while suboptimal in all Americans, are of particular concern in minority females. Little is known about cancer knowledge and screening behavior in Southeast Asian populations in the United States. We interviewed 38 Southeast Asian women of Cambodian or Vietnamese origin living in the Philadelphia, Pennsylvania, area. A telephone interview was conducted by bilingual/bicultural interviewers. Seventy-one percent (95% confidence interval [CI], 54% to 85%) of women in the study did not know what cancer was and 74% were unable to identify a cancer prevention strategy. Greater knowledge about cancer and identification of preventive measures were associated with employment outside the home, more years of education, and age, but not with length of time in the United States. Cancer education programs need to identify the patient's level of knowledge about cancer, elicit and respectfully address beliefs about causality and prevention, and ensure that health information is provided in a language understandable to the patient.

Vascular risk factors in cocaine users with stroke.

OBJECTIVES: To determine if the classic stroke risk factors, chronic hypertension, diabetes, cigarette smoking, chronic alcohol abuse, older age, and male sex, are also risk factors in cocaine-related ischemic and hemorrhagic stroke. METHODS: A computer search of ICD-9 codes identified 100 patients admitted to two inner city hospitals between 1986 and 1995 with acute ischemic or hemorrhagic stroke who also tested positive for cocaine in the urine. The case records of these patients were reviewed retrospectively. The ischemic and hemorrhagic stroke groups were compared with a control group of 109 cocaine users without a history of stroke. Multiple logistic regression was performed to see if the classic stroke risk factors independently increased the risk of cocaine-induced ischemic and/or hemorrhagic stroke. RESULTS: A total of 66 stroke patients in the study group had ischemic stroke, whereas 34 had hemorrhagic stroke. The stroke and control groups were similar in racial and gender composition. The mean ages of patients with ischemic and hemorrhagic stroke (both 41 years) were greater than the control group (34 years) (P<.01). Only chronic hypertension (odds ratio [OR] 5.2, P<.0001) and older age (OR 1.08/year increase of age, P<.0006) were independent risk factors for ischemic stroke. Female sex (OR 3.2, P<.015) and older age (OR 1.1/year increase of age, P<.0002) were independent risk factors for hemorrhagic stroke. CONCLUSIONS: Chronic hypertension and older age may magnify the risk of cocaine-related ischemic stroke, whereas female sex and older age may increase the risk of cocaine-related hemorrhagic stroke.

Are critically ill older patients treated differently than similarly ill younger patients?

Our goal was to determine whether critically ill older patients are treated differently than middle-aged patients. If so, what factors besides age contribute to that difference? Internal medicine residents (n = 46) and practicing internists (n = 41) received 8 clinical vignettes of 4 critically ill 85-year-old patients and 4 critically ill 50-year-old patients. Each patient had a distinct premorbid mental and physical state. Each respondent selected from 4 levels of therapeutic aggressiveness for each patient. The main outcome measure was the proportion of physicians who intended to treat the older of each matched pair of patients less aggressively than the younger one (that is, downgraded for age). Eight physicians (9%) treated a previously unimpaired 85-year-old patient less aggressively than a comparable 50-year-old patient. When the matched patients were either premorbidly mentally or physically impaired (but not both), about 20% of physicians downgraded for age. Most downgraded for age in matched patients who were premorbidly both mentally and physically impaired. We conclude that age alone does not engender much therapeutic bias against older patients as long as they are physically and mentally intact before the onset of their acute illness. As premorbid disabilities multiply, older patients may be treated less aggressively than younger ones with similar impairments and clinical presentations.

Inner-city achievers: who are they?

OBJECTIVE: To define characteristics that differentiate inner-city children with Average or above-Average Full Scale IQ scores (> or =90) from those with below-Average scores (<90). DESIGN: As part of a prospective study of children with and without in utero cocaine exposure tested at age 4 years on the Wechsler Preschool and Primary Scales of Intelligence-Revised, we found that, taken together, only 32 (21%) scored at or above 90 whereas 118 (79%) scored below 90. The groups (IQ> or =90 and IQ<90) were compared on prenatal, natal, and postnatal factors. SETTING: A study center in an inner-city hospital. PARTICIPANTS: One hundred-fifty children of low socioeconomic status, 34 weeks' gestational age or older, and nonasphyxiated at birth, who had intelligence testing at age 4 years; 150 caregivers (biological and foster). MAIN OUTCOME MEASURES: Association of Full Scale IQ with prenatal, natal, and postnatal characteristics (including caregiver-child interaction measured by the Parent Caregiver Involvement Scale [PCIS], and home environment measured by the Home Observation for Measurement of the Environment [HOME]). RESULTS: The group of children with IQs at or above 90 (n = 32) did not differ from the group with IQs below 90 (n = 118) in prenatal or natal characteristics (all P> or =.18) or proportion in foster care, attendance at day care or Head Start, continued caregiver cocaine use, or parental IQ. Children with IQs at or above 90 had more developmentally appropriate interaction by caregivers (P=.043) and higher scores on 6 of 8 subscales and Total HOME (P< or =.05) than the group of children with IQs below 90. CONCLUSIONS: Two postnatal factors, home environment and caregiver-child interaction, were associated with Full Scale IQ scores at or above 90 whereas prenatal and natal factors were not. These potentially malleable postnatal factors can be targeted for change to improve cognitive outcome of inner-city children.

Accuracy and clinical utility of the mini-dose 14C-urea breath test in the evaluation of Helicobacter pylori infection.

The aim of this study was to evaluate the accuracy of the 14C-urea breath test by comparing the results to those obtained by endoscopy with mucosal biopsy. We also examined the value of the breath test result obtained prior to endoscopy in predicting peptic ulcer disease. Forty-two individuals underwent the 14C-urea breath test. Collections of expired C02 were analysed using a liquid scintillation counter. All individuals then underwent endoscopy with biopsy. Biopsy material was evaluated by the rapid urease method and by histology for the presence of H. pylori. Our results demonstrated that the 14C-urea breath test was 100% sensitive and specific when compared to the rapid urease test as the 'gold standard' for the detection of H. pylori. In comparison to pathology, the sensitivity remained 100% and the specificity was 89%. The results of the 14C-urea breath test had a poor predictive value for the determination of peptic ulcer disease. We conclude that the 14C-urea breath test can be easily performed at any medical facility equipped with a liquid scintillation counter and can accurately detect H. pylori. A negative breath test result could not exclude the presence of peptic ulcer disease.

Children with in utero cocaine exposure do not differ from control subjects on intelligence testing.

OBJECTIVE: To determine if in utero cocaine exposure affects IQ scores in children at age 4 years. DESIGN: A prospective, longitudinal evaluation by blinded examiners of the IQ scores of cocaine-exposed and control children of low socioeconomic status who have been observed since birth. SETTING: A study center in an inner-city hospital. PARTICIPANTS: One hundred one children with in utero cocaine exposure and 118 control children, all of whom were 34 weeks' gestational age or older and nonasphyxiated at birth. MAIN OUTCOME MEASURE: Intelligence quotient scores on a standardized intelligence test, the Wechsler preschool and Primary Scale of Intelligence--Revised. RESULTS: Seventy-one cocaine-exposed and 78 control children were administered the Wechsler Preschool and Primary Scale of Intelligence--Revised. Maternal, natal, and 30-month characteristics of the children tested did not differ from those not tested. Groups did not differ on mean Performance (83.2 vs 87.0), Verbal (79.0 vs 80.8), or Full Scale (79.0 vs 81.9) IQ scores (all P > or = .10 [values for cocaine-exposed children given first]). None of these 3 scores was associated with cocaine exposure in multivariate linear regressions. Although cocaine-exposed and control groups did not differ in outcome, 93% of cocaine-exposed and 96% of control children had Full Scale IQ scores below 100, the mean IQ score for the test. CONCLUSIONS: In an inner-city cohort, IQ scores did not differ between cocaine-exposed and control children. However, both groups performed poorly.

Intracranial meningiomas: factors that influence the development of cerebral edema after stereotactic radiosurgery and radiation therapy.

PURPOSE: To evaluate causative factors of cerebral edema after stereotactic radiosurgery or stereotactic radiation therapy in intracranial meningiomas. MATERIALS AND METHODS: Of 43 adult patients with intracranial meningiomas, three received 13.5-18-Gy single-fraction stereotactic radiosurgery; one received 19.8 Gy in three fractions, one received 42 Gy in six fractions, and 31 received 32-36 Gy in six to eight fractions of stereotactic radiation therapy; and seven received 45-54-Gy external-beam radiation with 20-28 Gy in five to seven fractions as concomitant stereotactic boosts. Brain edema was estimated by calculating the edema index. RESULTS: After irradiation, all 11 patients with parasagittal and four patients with nonparasagittal tumors developed worsening cerebral edema that necessitated the administration of steroids (P < .001). The statistically significant factors for the development of edema were parasagittal location, presence of pretreatment edema, sagittal sinus occlusion, and the use of more than 6 Gy per fraction. Five patients with parasagittal tumors developed life-threatening panhemispheric edema, which was fatal in one. The causative factors of panhemispheric edema were a large tumor, single-fraction stereotactic radiosurgery, or use of more than 6 Gy per fraction. CONCLUSION: A smaller dose per fraction and aggressive use of steroids may help prevent life-threatening complications due to worsening edema.

Serum albumin is predictive of 30-day survival after percutaneous endoscopic gastrostomy.

BACKGROUND: Percutaneous endoscopic gastrostomy (PEG) tubes have been used since 1980 in patients who require prolonged enteral feeding. PURPOSE: To identify factors associated with poor survival (defined as survival < 30 days) post-PEG. METHODS: We assessed a cohort of 64 patients consecutively referred for a percutaneous endoscopic gastrostomy tube in a single, tertiary care hospital. Prior to PEG tube placement, we evaluated relevant clinical variables in each patient to determine their effect on 30-day mortality. RESULTS: Of the cohort, 43 of 64 (67.2%) survived at least 30 days after tube placement. One death was attributable to tube placement. Bivariate analyses showed that 30-day survival correlated directly with serum albumin (r = .253; p = .049) and inversely with creatinine (r = -0.255; p = .042). Using multivariable logistic regression analysis, only albumin was identified as an independent predictor of 30-day survival (p = .044). Eighty-three percent of patients with a serum albumin > or = 3.0 g/dL survived 30 days compared with 58% with an albumin < 3.0, a difference of 25% (95% CI, -2% to 54%; p = .07). CONCLUSIONS: In conclusion, serum albumin appears to be a predictor of early survival in individuals undergoing percutaneous endoscopic gastrostomy tube placement.

Alkalinization is ineffective for severe hyperkalemia in nonnephrectomized dogs. Hyperkalemia Research Group.

OBJECTIVE: To determine whether alkalinization with sodium bicarbonate (NaHCO3) in near-lethal hyperkalemia either lowers potassium (K) rapidly or shortens duration of cardiac conduction disturbances. METHODS: A controlled canine laboratory investigation of 3 treatments for severe hyperkalemia. Conditioned dogs (n = 8; 17-30 kg) received, in random order, 2 mmol/kg of each of 3 treatments (matched in sodium and water) in separate experiments > or = 1 week apart: 1.05% NaHCO3 over 60 minutes (infusion therapy); 8.4% NaHCO3 over 5 minutes, then 14 mL/kg sterile water over 55 minutes (bolus therapy); 8.4% NaCl over 5 minutes, then 14 mL/kg sterile water over 55 minutes (saline therapy). Prior to administering one of the above therapies, the animals were anesthetized with 0.5-2.5% isoflurane and ventilated to maintain a normal PCO2. After 30 minutes of equilibration, 2 mmol/kg/hr (loading dose) of a 2-mmol/mL KCl solution was given until idioventricular or relative junctional bradycardic dysrhythmias were sustained for 15 minutes. Then KCl was decreased to 1 mmol/kg/hr (maintenance dose) for 2 hours and 45 minutes. Treatment was begun after 45 minutes of maintenance KCl infusion. RESULTS: The pretreatment K level (all studies) was 9.06 +/- 0.82 mmol/L (mean +/- SD). Although the mean K level decreased more after saline therapy than after bolus therapy at every time, differences were neither statistically significant nor clinically important during the first 30 minutes. The means of the differences in decreases (saline minus bolus) were small, 0.26 (95% CI, -0.48 to 1.00) at 15 minutes, 0.16 (95% CI, -0.67 to 0.98) at 30 minutes. Dysrhythmia duration was shorter with bolus therapy than for saline therapy in only 1 of 5 dogs (p = 0.38). CONCLUSIONS: Hypertonic saline bolus lowered plasma K as effectively as NaHCO3 bolus in this animal model within the first 30 minutes. Clinically meaningful decreases due to alkalinization alone within 30 minutes are unlikely.

Play behavior in toddlers with in utero cocaine exposure: a prospective, masked, controlled study.

Play behavior was compared between toddlers with in utero cocaine exposure and controls of similar low socioeconomic status enrolled at birth in a prospective, masked study. At 18 and 24 months, 83 cocaine-exposed and 93 control toddlers were videotaped playing on their own for 15 minutes. An observer who was off-site and unaware of project purpose or drug exposure status of toddlers recorded the most cognitively complex play activity per 15-second interval. In a total of 315 play sessions, the groups did not differ in middle and highest level of play achieved at either 18 or 24 months (p > or = .27). After controlling for confounders, the proportions of play behavior in each of six play categories were similar in the two groups at both 18 and 24 months (p > or = .42). We conclude that in utero cocaine exposure was not associated with differences in play behavior in this cohort of cocaine-exposed and control toddlers.

Predicting clinical states in individual patients.

A probability model expresses the relation between the presence of clinical findings (input or independent variables) and the probability that a clinical state will occur (the dependent variable); for example, it expresses the probability that a disease is present or will develop or the probability that an outcome state will be reached. Probability models are developed by using selected study groups. Although these models are most often used to make predictions for groups of patients, they can also predict clinical states for individual patients. The following seven criteria provide a basis for the critical appraisal of probability models. In particular, physicians can use these criteria to decide when a specific probability model should be used to make a prediction in an individual patient. Five of the criteria are concerned with the applicability of a model to a particular patient: 1) the comparability of the patient and the study group used to develop the model; 2) the congruence between the clinical state of interest to patient and physician and the model's outcome; 3) the availability of all input variables where and when the prediction is to be made; 4) the usefulness of a quantitative estimate of the predicted clinical state; and 5) the degree of uncertainty in the probability estimate. The other two criteria are concerned with how well the probability model "works": 6) the fit of probabilities calculated from the model to the outcomes actually observed and 7) the model's ability to discriminate between outcome states relative to chance and to other, more traditional, prediction methods. We illustrate the use of these criteria by applying them, in the form of questions, to a convenient, tabular version of a model that estimates a patient's chances of surviving for 10 years after having definitive surgical therapy for primary cutaneous melanoma.

Natal status of infants of cocaine users and control subjects: a prospective comparison.

Cocaine use by pregnant women has increased dramatically in recent years. To assess the effect of maternal cocaine use on infant outcome, we enrolled 224 women (105 cocaine users, 119 control subjects) and their infants (all of 34 weeks or more gestational age and nonasphyxiated) in a prospective, blinded study. Results showed that infants exposed to cocaine were more likely to be admitted to the newborn intensive care unit, be treated for congenital syphilis, have a greater length of stay, and be discharged to a person other than the mother (all p < 0.01). Birth weight and head circumference, adjusted for gestational age, were smaller in the infants exposed to cocaine than in control infants (p < 0.001). After statistically controlling for cigarette use and other confounders, however, the odds of infants exposed to cocaine and control infants having birth weight and head circumferences less than the 25th percentile for gestational age did not differ (both p > 0.80). Infants of cocaine-using mothers and control subjects had a similar incidence of abnormal cranial and renal ultrasonographic findings and abnormal pneumocardiograms (all p > or = 0.32). We conclude that in this cohort of nonasphyxiated infants of 34 weeks or more gestational age, infants exposed to cocaine had more medical and social problems than control infants but did not differ statistically in the incidence of severe growth retardation, abnormal cranial or renal ultrasonographic findings, or abnormal pneumocardiograms. We suggest that natal interventions for the nonasphyxiated term and near-term infant exposed to cocaine should include a careful history and physical examination, follow-up plans, and social service involvement.