Retrospective Study of Symptoms in Post-Discharge Patients
Epstein K, Juarez E, Loya K, et al. Frequency of new or worsening symptoms in the posthospitalization period. J Hosp Med. March/April 2007;2(2):58-68.
As hospital stays shorten and acuity rises, patients often are discharged with complex instructions and discharge plans including home health services, physical therapy, hospice service, antibiotic infusions, and follow-up appointments. The potential for new or progressive symptoms in the days following discharge is an important parameter in assessing whether our planning is safe and effective.
The researchers in this study investigated the post-discharge period using a retrospective analysis of new or worsening symptoms within two to five days of hospital discharge among 15,767 patients surveyed between May 1 and Oct. 31, 2003. Patients were all under the care of hospitalists employed by IPC, a large private hospitalist group based in North Hollywood, Calif. Total discharges from which this cohort was selected numbered 48,236.
Staff with medical backgrounds conducted a scripted survey by phone. Licensed nursing personnel contacted those patients whose answers to initial questions suggested they were at high risk for postdischarge complications. A five-point Likert scale was used so patients could rate their overall health status in addition to specific symptomatology ranging from abdominal pain to bleeding. Other questions targeted pick-up and administration of prescribed medications, insulin regimen adherence, and implementation of home health services.
Among all patients discharged, 32.7% were contacted within two days of discharge. The mean age was 60.1 years, and 57% were female. Ethnicity and socioeconomic status were not reported. Medicare and HMOs were the most common type of insurance. Of the 15,767 patients contacted, 11.9% reported symptoms that were new or worsening since discharge; of this subgroup, 64% had new symptoms whereas 36% had “worse” symptoms.
Women were more likely than men to report new or worsening symptoms, and patients who rated themselves as having a poor health status were more likely to have new or worsening symptoms. Younger patients were less likely to report new or worsening symptoms, particularly younger men. Those with new or worse symptoms were slightly more likely to have made a follow-up appointment but also more likely to have a problem with their medications. Interestingly, there was no correlation between self-rated health status and reported severity of illness based on the diagnosis related group (DRG) score. Patients discharged with a DRG of chest pain were less likely to report symptoms than all other patients.
The authors acknowledge the low response rate (32.7%) relative to the 48,236 discharges during the study period. Logistic challenges, resource limitations, and erroneous contact information precluded successful contact for the remainder of patients. The magnitude of this exclusion effect essentially precludes statistically valid extrapolation to the inception cohort (all discharges). For example, in a sensitivity analysis where all the excluded patients are assumed to have developed new or worsening symptoms, the actual rate overall would have been 71%. If none developed new or worsening symptoms, that rate would be 3.8%. The rate for the inception cohort may or may not approximate the 11.9% found among the studied patients. There is insufficient evidence to determine whether the studied cohort reflects the entire population of discharged patients.
To their credit, no such analysis or interpretation is claimed or intended by the authors, and the information derived from the included cohort nonetheless provides interesting and important descriptive data.
Ethnicity and cultural factors were not taken into consideration. One might postulate that language barriers could affect compliance and symptom reporting. Day-of-the-week and holiday status also were not reported with regard to discharge. It would be interesting and useful to know whether access to pharmacy and other resources varied in this regard and whether symptom reporting was affected by such timing.
In the final analysis, this study suggests hospitalists remain alert to possible problems that might develop during the vulnerable first few days following discharge. It reminds us to advise patients how to receive prompt and knowledgeable medical advice from someone familiar with their hospital care prior to their first scheduled follow-up.
Based on the reported rate of new or worsening symptoms, should a post-discharge clinic be part of hospitalists’ scope of practice, at least for selected patients? Can subsets of patients who would benefit most from such intervention be identified? These and many more questions are raised by this study. We look forward to further research into the best process for ensuring optimal outcomes in the immediate post-discharge period.
Rosiglitazone’s Effect on MI Risk in Diabetes Patients
Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med. 2007 June 14;356(24):2457-2471.
Cardiovascular causes account for more than 65% of deaths in diabetic patients. Rosiglitazone—a thiazolidinedione-class drug—has been broadly used in diabetes, but its effect on cardiovascular morbidity and mortality has not been conclusively determined. The authors initiated this meta-analysis to determine the effect of rosiglitazone on the risk of myocardial infarction (MI) and death from cardiovascular causes in diabetics.
The meta-analysis included 42 trials from three data sources. Forty trials were obtained from the Food and Drug Administration (FDA) Web site and the GlaxoSmithKline clinical trials registry. The third data source comprised two recent large, well-known trials: the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) and the A Diabetes Outcome Prevention Trial (ADOPT).1-2 The authors’ inclusion criteria were a study with a duration of more than 24 weeks, the use of a randomized control group not receiving rosiglitazone (placebo or comparator drug), and the availability of outcome data for MI and death from cardiovascular causes.
The studies included 15,560 patients randomly assigned to regimens that included rosiglitazone and 12,283 patients to comparator groups that did not include rosiglitazone.
The authors reviewed the data summaries of the 42 trials and tabulated adverse events (not reported as outcomes) of MI and death from cardiovascular causes. Hazard ratios could not be calculated since time-to-event data were lacking. Summary data also precluded the ability to determine whether the same patient suffered both an MI and death from cardiovascular causes.
Results of the authors’ statistical analyses included odds ratios and 95% confidence intervals to assess the risk associated with the rosiglitazone group as well as the subgroups of metformin, sulfonylurea, insulin, and placebo versus rosiglitazone.
The authors tabulated 86 MIs and 39 unadjudicated deaths from cardiovascular causes in the rosiglitazone group, and 72 MIs plus 22 deaths from cardiovascular causes in the control group.
The main conclusion was that rosiglitazone was associated with a statistically significant increase in the risk of MI (odds ratio 1.43, 95% confidence interval 1.03 to 1.98, p=0.03), but was not associated with a statistically significant increase in the risk of death from cardiovascular causes (odds ratio 1.64, 95% confidence interval 0.98 to 2.74, p=0.06).
Additionally, there were no statistical differences between rosiglitazone versus placebo or the individual antidiabetics in the subanalyses.
The authors have recognized the following major limitations in this meta-analysis:
- The low rate of MI is 0.55% (86 of 15,560 cases) in the rosiglitazone group and 0.59% (72 of 12,283 cases) in the control group. The odds ratio of 1.43 was statistically significant in the rosiglitazone group, although the event rate was higher in the control group. The risk of cardiovascular death was not significant, though a trend toward a higher death rate is noted;
- The lack of source data did not allow the use of time event analysis including hazard ratios;
- The definition of MI was unavailable; and
- MI and cardiovascular events were recorded in the trials as adverse events, not outcomes. Therefore, deaths from the latter were unadjudicated.
The authors suggested that the potential mechanism for increased MI in the rosiglitazone group could be its known effects on increasing low-density lipoproteins (LDL), precipitating congestive heart failure and reducing hemoglobin levels.
Rosiglitazone is one of two peroxisome proliferation activated receptor y (PPAR-y) agonists licensed for use in the United States; the other is pioglitazone. The third drug was troglitazone; it was taken off the market in March of 2000 due to hepatotoxicity.
The PPAR-y agonists decrease plasma glycemia by increasing insulin sensitivity in the peripheral tissues. These drugs have complex physiologic effects in activating and suppressing multiple genes, with most target genes being unknown. The observed side effects with rosiglitazone are not necessarily a class effect. Pioglitazone showed a trend toward reducing triglycerides and cardiovascular events, including MI and CVA, in a prospective, randomized trial called Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROACTIVE).
This meta-analysis precipitated an interim analysis of the ongoing Rosiglitazone Evaluated for Cardiovascular Outcomes and Regulation of Glycemia in Diabetes (RECORD) trial.3 The RECORD trial is a randomized, open-label, multicenter, non-inferiority trial of 4,427 patients; 2,220 received add-on rosiglitazone, and 2,227 received a combination of metformin plus sulfonylurea (control group). The primary end point was hospitalization or death from cardiovascular causes. Interim findings were inconclusive for the rosiglitazone group. There was also no evidence of any increase in death from cardiovascular causes or all causes. However, rosiglitazone was found to be associated with an increased risk of congestive heart failure. The data were insufficient to determine whether the drug was associated with increased MI risk.
This important meta-analysis raises concerns about the association of rosiglitazone with cardiovascular events—but do not consider it definitive. For now, patients with comparable alternatives to rosiglitazone (indeed all patients on this medication) should be advised of the undetermined safety concerns. For those who consider rosiglitazone a compelling choice, abrupt discontinuation on the basis of this study may be premature.
Finally, we need to remain cognizant of the proven negative side effects of rosiglitazone—it increases fracture risks in women, precipitates congestive heart failure, increases LDL, and decreases hemoglobin levels. We should consider alternative anti-hyperglycemic agents in selected patients at risk until there are solid data from large randomized control trials with rosiglitazone that pre-empt its use altogether.
- Gerstein HC, Yusuf S, Bosch J, et al. Effect of rosiglitazone on the frequency of diabetes in patients with impaired glucose tolerance or impaired fasting glucose: a randomized controlled trial. Lancet 2006 Sep 23; 368(9547):1096-1105.
- Kahn SE, Haffner SM, Heise MA, et al; ADOPT Study Group. Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy. N Engl J Med. 2006 Dec 7;355(23):2427-2443.
- Home PD, Pocock SJ, Beck-Nielsen H, et al. Rosiglitazone evaluated for cardiac outcomes and regulation of glycemia in diabetes (RECORD): study design and protocol. Diabetologia. 2005;48:1726-1735.
Statins and Sepsis in Dialysis Patients
Gupta R, Plantinga LC, Fink NE, et al. Statin use and hospitalization for sepsis in patients with chronic kidney disease. JAMA. 2007 Apr 4;297(13):1455-1464.
Epidemiological data has revealed an increase in the rate of sepsis in the U.S. during the past two decades.1 In individuals with chronic kidney disease who are on dialysis, sepsis is a significant cause of morbidity and mortality. Various studies have looked at risk factors associated with septicemia in patients with chronic kidney disease; however, no preventive treatments have been identified.
Recent research has shown the use of statins has been associated with a decreased rate of sepsis and improved sepsis outcomes. The authors of this study investigated whether statin use may help reduce the incidence of sepsis in patients with chronic kidney disease on dialysis.
This prospective cohort study enrolled 1,041 participants attending dialysis clinics from October 1995 to June 1998, with a follow-up through Jan. 1, 2005. Statin use at baseline was determined by review of medical records. The primary outcome was hospitalization for sepsis, indicated by hospital data from the U.S. Renal Data System (mean follow-up 3.4 years).
The association of statin use and sepsis was assessed using two analyses. A multivariate regression analysis was performed on the entire cohort, and adjustments were made for potential confounders. An analysis was performed on a sub-cohort comparing sepsis rates in statin users with a control group identified through the likelihood of having been prescribed a statin (propensity matching).
There were 303 hospitalizations for sepsis among the 1,041 patients enrolled, with 14% of participants receiving a statin at baseline. The crude incidence rate of sepsis was 41/1,000 patient-years among statin users compared with 110/1,000 patient-years in the control group (p<0.001). The fully adjusted incidence ratio for sepsis among statin users versus nonusers was 0.38, or 62% lower among statin users.
In the propensity-matched subcohort group, there were 54 hospitalizations during follow-up. The relative risk of sepsis was 0.24 (95% confidence interval, 0.11-0.49) for statin users compared with nonusers.
A strong and independent association exists between statin use and reduced incidence of sepsis in chronic kidney disease patients. This association remained statistically significant after controlling for potential confounding. Why the statins might have this effect is not definitively known.
This national study further demonstrates the potential protective effect of statins on the occurrence of sepsis, which has been observed in previous research in a non-renal population. The author mentions that this is the first study to show a strong and significant effect of a medication administered long term on lower rates of sepsis among patients with chronic kidney disease.
Because this is an observational study, it is limited due to lack of randomization. As such, this study cannot prove causality. Further limitations include the assessment of patient and treatment factors at baseline, which can lead to a misclassification of factors that change over time. It is important to point out the study was dependent on U.S. Renal Data System and Medicare data to determine outcome, and the use of their ICD-9 coding information may have resulted in decreased reporting of sepsis.
Still, the relevant results of this investigation warrant further examination of statins and the prevention of sepsis in a prospective randomized trial. TH
- Sarnak MJ, Jaber BL. Mortality caused by sepsis in patients with end-stage renal disease compared with the general population. Kidney Int. 2000 Oct;58(4):1758-1764.