## Pho.sphhp.buffalo.edu

**University at Buffalo **
**The State University of New York **
School of Public Health and Health Professions

*“Dedicated to improving health through population based research.” *
**Acknowledgement **
**Primary Study Investigator **
**Contributors **
This paper focuses on the relationship between hypertension and non-steroidal anti-
inflammatory medications. Research has indicated that non-steroidal anti-inflammatory drugs
increase blood pressure, and in some cases the magnitude of the increase is clinically significant.
The data used for this paper are from the National Center for Health Statistics. The data from the
National Hospital Ambulatory Medical Care Survey and the National Ambulatory Medical Care
Survey were downloaded from the CDC’s website, and concatenated as they both contained the
necessary information. Information regarding these data sets is available a
/about/major/ahcd/ahcd1.htm. Both data sets used a four-stage probability sample, and the fields
of interest were demographics, diagnoses, medications prescribed, and blood pressure (systolic
The background for this paper is that research has shown that non-steroidal anti-
inflammatory medications increase blood pressure. Particular classes of NSAIDS of interest are
ibuprofen, naproxem, toradol, celecoxib, and rofecoxib. This paper attempts to answer the
question of how NSAIDS and a diagnosis of hypertension, and the interaction between these two
factors, affect systolic and diastolic blood pressure.
The NHAMCS and NAMCS data set were downloaded from the NCHS’s website. Both
the outpatient and emergency data sets were used from the NHAMCS survey, as the instructions
in regards to sample design instructed. The software package SAS was used in this analysis.
The procedure PROC SURVEYREG was used to perform the analysis, which is a linear
regression procedure that allows for the inclusion of sample design variables. The sample design
variables for cluster (cpsum), for strata (cstratm), and for weight (patwt) were included in the
data sets, and were appropriately included in the SAS procedure, documentation is available at
the siteThe actual linear model fitted was:

*BP * 0 1*

*HTN * 2 *

*NSAID * 3* (

*HTN **

*NSAID*) 4 *

*RACE * 5 *

*SEX * 6 *

*AGE*
where terms are defined in the following table and subsequent paragraph.
Also, note that HTN stands for a diagnosis of hypertension, NSAID refers to an
individual being prescribed one or more NSAID medications, HTN*NSAID refer to the four
possible combinations of the previous two variables, the RACE variable has four levels (white,
African-American, Hispanic, and other), the sex variable has two levels, and age is integer
valued but is considered in the model as continuous.
This model was fitted for both systolic and diastolic blood pressure. The analyses
involved computing average least square estimates for blood pressure given distinct
combinations of the HTN and NSAID variables. The method for computing the estimates
involved the summation of the predicted value of blood pressure for each distinct combination of
race, sex, and age and then division by the total number of combinations as follows :
0 1*

*HTN * 2*

*NSAID * 3*(

*HTN **

*NSAID*) 4*

*RACE * 5*

*SEX * 6*

*AGE*
and then this quantity is divided by the total number of summations performed (note that a
represents age, s represents sex, r represents race. In terms of the types of the variables, the
In regard to both the model for systolic and diastolic blood pressure, all of the predictor
variables have significant p values (<0.055), except for the interaction term, in both cases, and
NSAID for DBP. We can conclude that while the NSAID variable and HTN variable have a
significant effect on blood pressure, the interaction between these two factors does not
significantly affect blood pressure. In other words the extent to which non-steroidal medications
increase blood pressure is statistically similar in both the hypertension and non-hypertension
groups. The following graphs visually illustrate this relationship.
Table 2 : Model Effects for Diastolic Blood Pressure
Table 3 : Model Effects for Systolic Blood Pressure
Table 4: Parameter Estimates for Diastolic Model
Table 4 : Parameter Estimates for Systolic Model
Graph 1: Interaction Effect for Diastolic Model:
Graph 2 : Interaction Effect for Systolic Model :
This analysis argues that the effect of non-steroidal anti-inflammatory medications is
independent of a diagnosis of hypertension. That is the effect of NSAIDS on systolic blood
pressure is essentially the same in both the hypertension and non-hypertension groups. The
analysis, however, indicates that non-steroidal anti-inflammatory drugs do significantly increase
systolic blood pressure. These results suggest that NSAID usage is associated with increase
systolic blood pressure independent of hypertensive status.

Source: http://pho.sphhp.buffalo.edu/training/students/reports/NSAID_report.pdf

Die Wissenschaft ist ständig im Fluss. Die nachfolgenden Ehrlich gesagt, uns hat dieses Ergebnis wenig überrascht: Denn vor lauter Skandalen um Informationen beruhen auf gründlichen Recherchen des die neuen Mittel hat sich einfach niemand über die älteren herkömmlichen Präparate Ge-Herausgebers. Das Ziel der Publikation ist es, diese mo-danken gemacht. Die Erfahrung zeigt, dass nebenw

PRESSE-INFORMATION Zivilisationskrankheit Osteoporose - Neue wissenschaftliche Erkenntnisse: Gattung ‚Montmorency’- Kirschen verbessern natürlich die Knochenstabilität Schätzungsweise gibt es in Deutschland insgesamt 6,3 Mio. Menschen, die nach dem 50. Lebensjahr unter Osteoporose leiden - ca. 83% davon sind Frauen. Die Dunkelziffer der Osteoporose Erkrankten dürfte noch wei