February 1997
Pain and satisfaction with pain control in seriously ill hospitalized adults: Findings from the SUPPORT research investigations

Desbiens NA, Wu AW, Broste SK, Wegner NS, Connors AF, Lynn J, Yasui Y, Philips RS, Fulkerson W;

Crit Care Med 1996; 24:1953-1961

[ see abstract below ]

In the December 1996 issue of Critical Care Medicine, Desbiens et al. report on the incidence and severity of pain in patients enrolled in SUPPORT (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments). Readers may recall the article published by the SUPPORT group in JAMA last Fall which suggested that the use of the pulmonary artery catheter was associated with a higher mortality in this population (Connors AF, Jr., Speroff T, Dawson NV, et al. -- The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276:889-97; Dr. Lubarsky reviewed this article in the December edition of AnesthesiaWeb. The Abstract is also here.

There are two reasons why this article might be relevant to anesthesiologists. First, it attempts to define the incidence and severity of pain in critically ill patients - and as experts in pain control, anesthesiologists should be at the forefront of analgesic therapy in the ICU. Second, it provides further insight into the scope and limitations of SUPPORT, which is having such an extraordinary impact on our view of pulmonary artery catheterization.

In SUPPORT 9,105 patients were studied in five hospitals over a four year period. The patients were carefully selected to have short life expectancy with "high mortality" disease. This included acute or chronic respiratory failure, congestive heart failure, chronic liver failure, nontraumatic coma, specific forms of cancer (metastatic colon cancer and advanced small cell lung cancer), and multisystem organ failure from malignancy or sepsis.

It specifically excluded the trauma and general surgical population. Also excluded were patients with AIDS, those who died within 48 hours of ICU admission, who were scheduled for discharge within 72 hours, who were pregnant or who did not speak English.

The first two years was an observational study. During the second two years, nurse clinicians provided physicians with prognostic and preference reports. The original intent of SUPPORT was to define the impact of ongoing prognostic data on physician management and patient outcome. From the huge amount of data obtained, there have been a number of "spin off" studies, of which this is one.

Patients (or their "surrogates" - relatives who would have provided information about patients unable to speak) were interviewed on day 2 and 8 in the ICU regarding the frequency and severity of their pain, and whether they were satisfied with its treatment. If the patient was not able to answer the investigators "used surrogate responses calibrated to patient responses" (sic).

The investigators were able to interview 3,571 patients and 1,605 surrogates (by day 8, 16% had already died, 17% were already discharged, and 23% were unable to communicate). The results revealed that nearly 50% of patients had pain some of the time, and 15% had severe pain at least half the time. In 15% of these (i.e. about 2% of the total), dissatisfaction was expressed about the quality of pain control.

There was a positive relationship between pain and surgical patients, metastatic cancer, and dependency, comorbidity, anxiety, depression and poor quality of life. Older, sicker patients appeared to complain less of pain.

The authors concluded that "Better pain management strategies are needed for patients with the serious and common illnesses studied in SUPPORT".

What should we conclude from this? First, let's examine the limitations of this study, because there are many. The study selection specifically excluded the vast majority of patients that anesthesiologists routinely take care of in the OR (which should also be kept in mind in interpreting the findings regarding pulmonary artery catheters in the JAMA article).

It could be assumed that in the general postsurgical and posttrauma population pain will occur in 100% of patients - perhaps we need better data on its severity and success of treatment in this group as well. Perhaps, too, it could be salutary to consider the relatively high incidence of pain in the "medical" population studied here - a fruitful area for anesthesiologists to ply their trade!

The results are based on two interviews which did not distinguish between acute or chronic pain, or the effect of analgesia. Some of the data seem far-fetched - for example, it is reported that 35% of patients with nontraumatic coma had pain (?), and the statistical techniques used to equate surrogate and patient responses reek of "spin" to this reviewer! At one point the authors state: "The only reliable way to know about patients' pain is to ask them". Aha! There's the rub!

Most intubated, sedated, confused and disoriented ICU patients will not be able to express their feelings of pain! Perhaps anesthesiologists have something to teach our internal medicine colleagues about the recognition of the signs of inadequate anesthesia or analgesia (tachycardia, hypertension, lacrimation, sweating, pupillary dilation). It is likely, therefore, that the true incidence of pain and discomfort in the population studied may have been substantially higher.

An accompanying editorial sheds further light on the limitations of the study. In particular, it criticizes the superficial and nonstandardized methods of pain evaluation in an homogenized population (all of whom were English speaking and 80% white), which leaves "many pressing questions regarding pain and its control ... unanswered".

At the risk of being cynical, one wonders how many of these limitations are due to retrospective analysis by the authors of the SUPPORT data, which appeared to have originally had a quite different hypothesis and intent. Nonetheless, it does suggest that critical care anesthesiologists have a great deal to contribute both clinically and by means of careful investigation in the arena of pain and its control in the sick medical population.

Return to the Current Literature Review Front Page, or read the abstract:

ABSTRACT