key: cord-0006452-o6b6dcgd
authors: Farina, Maria Luisa; Bonati, Maurizio; Iapichino, Gaetano; Pesenti, Antonio; Procaccio, Francesco; Boselli, Luigi; Langer, Martin; Graziina, Augusta; Tognoni, Gianni
title: Clinical Pharmacological and Therapeutic Considerations in General Intensive Care: A Review
date: 2012-10-23
journal: Drugs
DOI: 10.2165/00003495-198734060-00003
sha: 0231e91dbdf6a9e93b6816f535331328e54982ed
doc_id: 6452
cord_uid: o6b6dcgd

The application of clinical pharmacological concepts and therapeutic standards in intensive care settings presents particularly difficult problems due to the lack of adequately controlled background information and the highly variable and rapidly evolving clinical conditions where drugs must be administered and their impact evaluated. In this review, an attempt has been made to discuss the available knowledge within the framework of a problem-oriented approach, which appears to provide a more clinically useful insight than a drug-centred review. Following a brief discussion of the scanty data and the most interesting models to which reference can be made from a pharmacokinetic point of view (the burn patient being taken as an example), the review concentrates on the main general intervention strategies in intensive care patients. These are based mainly on non-pharmacological measures (correction of fluid and electrolyte balance, total parenteral nutrition, enterai nutrition, oxygenation and ventilatory management) and are discussed with respect to the specific challenge they present in various clinical conditions and organ failure situations. In addition, 4 major selected clinical conditions where general management criteria and careful use of prophylactic and therapeutic drug treatments must interact to cope with the variety of presentations and problems are reviewed. These include: acute cerebral damage; anti-infective prophylaxis and therapy; cardiovascular emergencies; and problems of haemostasis. Each problem is analysed in such a way as to frame the pharmacological intervention in its broader context of the underlying (established or hypothesised) pathophysiology, with special attention being paid to those methodological issues which allow an appreciation of the degree of reliability of the data and the recommendations which appear to be practiced (often haphazardly) in intensive care units. The thorough review of the published literature provided (up to mid-1986) clearly shows that in this field the quality of randomised controlled and epidemiological studies is rather unsatisfactory. It would be highly beneficial to research and to clinical care if larger multicentric protocols and prospective epidemiological comparative investigations could be carried out to investigate more timely and adequately the variables which determine drug action, and the final outcome in the many subgroups of patients which must be considered in a proper stratification of intensive care unit populations.

of prophylactic and therapeutic drug treatments must interact to cope with the variety of presentations and problems are reviewed. These include: acute cerebral damage; anti-infective prophylaxis and therapy; cardiovascular emergencies: and problems of haemostasis. Each problem is analysed in such a way as to frame the pharmacological intervention in its broader context of the underlying (established or hypothesised) pathophysiology, with special attention being paid to those methodological issues which allow an appreciation of the degree of reliability of the data and the recommendations which appear to be practiced (often haphazardly) in intensive care units. The thorough review of the published literature provided (up to mid-1986) clearly shows that in this field the quality of randomised controlled and epidemiological studies is rather unsatisfactory.

It would be highly beneficial to research and to clinical care if larger multicentric protocols and prospective epidemiological comparative investigations could be carried out to investigate more timely and adequately the variables which determine drug action, and the final outcome in the many subgroups of patients which must be considered in a proper stratification of intensive care unit populations.

This article has arisen from a series of joint research programmes of a group of clinicians working in general and highly specialised intensive care units, and a group of clinical pharmacologists interested in the assessment of what in their discipline really does matter in clinical practice (Bonati & Tognoni 1984; Farina et al. 1981; Tognoni et al. 1980) . From a clinical pharmacological point of view, intensive care may be seen as a situation where contradictory situations and facts are frequently met: for example, the use of many drug treatments primarily tested under non-intensive care conditions; the lack offormal evaluation of the interplay between pharmacological and non-pharmacological interventions which are often utilised side by side and not as components of an integrated management approach; the call for a better defined pharmacokinetic profile of drugs which are given mostly at constant dosage rates while the vital functions and the body compartments of the patient are rapidly changing; and the difficult challenge of defining standard conditions of care against which the specific role of a new intervention can be comparatively tested.

Following other attempts (Chernow & Raymond 1983; Majerus 1982) , a systematic discussion of the problem is proposed with 3 main aims:

1. Consideration of the intensive care patient as one whose management needs most often take precedence over any search for a specific benefit to be derived from a particular pharmacological treatment: drugs are but one of the variables to be considered while pursuing a rational management programme.

2. Adoption of an approach that is tailored to the real-life situation of intensive care where patients are progressively and tentatively evaluated as their status is evolving, rather than assigned to a well-defined and stable diagnostic category. From this perspective, an approach where drugs are seen as specific tools aiming at specific targets loses its strength in favour of an attitude where active adjustment of therapeutic decisions and continuous evaluation are mandatory. This interplay is of course required in many other clinical situations; however, its importance in intensive care is underlined by the urgency of the situation and the uncertainty or unavailability of many pieces of information.

3. Consideration of intensive care as a situation where straight answers and guidelines must be accompanied by continual questioning and re-evaluation of the widely open methodological problems that exist behind 'recommended' treatments.

The rationale for selecting the topics to be discussed in this review lies in the reality of general intensive care units (ICUs) [Abizanda Campos et al. 1980; Farina et al. 1981; Merriman 1981] . Intensive perinatal or coronary care problems will not Glossary of terms ARDS Acute respiratory distress syndrome: an acute respiratory failure and distress associated with a specific incident or illness with the exclusion of exacerbation of chronic lung disease (National Institutes of Health 1972) FI~ Inspired oxygen fraction. Oxygen in high concentration is required to maintain life during certain lung diseases, but it carries the risk of toxicity (Deneke & Fanburg 1982) CMV Controlled mechanical ventilation: a form of respiratory support in which gases are pumped into the patient's lung by a mechanical ventilator. The expiration is controlled by the ventilator but is most often passive Paw Mean airways pressure (averaged for time during the respiratory cycle)

Positive end-expiratory pressure: airways pressure is not allowed to decrease below a set value during expiration (Shapiro et al. 1983; Weisman et al. 1982) IRV Inverse ratiO ventilation: a form of mechanical ventilation. Inspiration is greatly prolonged, giving high inspiratory time ratios (e.g. 4: 1) [Baum et al. 1980] IMV Intermittent mandatory ventilation: a form of mechanical ventilation in which spontaneous breathing is permitted but supplemented by a limited number of mechanical breaths (Weisman et al. 1983) HFPPV High frequency positive pressure ventilation: a form of respiratory support in which the ventilator provides high ventilator frequencies (from 60/min to co) at relatively low tidal volumes (Sjostrand & Eriksson 1980) Differential lung ventilation A form of mechanical ventilation by which each of the two lungs is ventilated independently according to its own physiological needs (Powner et al. 1977) FRC Volume of gas remaining in the lungs at end-expiration TSlC Total static lung compliance: the amount of gases that can enter the lung generating a unit change in pressure -a measure inversely related to lung stiffness VD!VT Physiological dead space: the portion of tidal volume that does not participate in gas exchange Barotrauma Damage to the lung due or related to an increase in airway pressure (e.g. pneumothorax); sometimes extended to the damage produced by high airways pressure on organs other than the lung (e.g. liver and kidney function impairment) [Johnson & Hedley Whyte 1972; Kumar et al. 1973] 664 be considered because the situations which are met in these settings are much better defined, provide a relatively easier ground for a positive interaction of clinical pharmacology and clinical care, and the literature related to specific drug therapies and clinical conditions is well developed and is covered periodically in careful reviews.

The relationship of drug dosing and disposition to drug effects, efficacy and toxicity has been extensively investigated and documented and is readily available in comprehensive and easily accessible reviews, which are regularly updated (Gibaldi & Prescott 1983) . From a systematic scrutiny of specialty journals in this area, it is easy to verify that while the information is abundant and often detailed for disease states where a single organ or system is impaired, the data on clinical conditions where multiple organ failure is the rule are rather scanty. This is also true with respect to prognostic and descriptive mathematical modelling. Such a finding is not unexpected, because of the obvious difficulties that are inherent in studies where animal models mimicking complex clinical conditions are not readily available (Bortolotti & Bonati 1985) and which would require a close and long term interaction of clinical pharmacologists with intensive care clinicians to produce enough data to represent reliably the reality of intensive care patients or populations.

Possibly the most comprehensive model for what could be an optimum approach to complex situations is described in a series of publications on bums patients, where cardiovascular, hepatic, renal and dermatological functions are contemporaneously involved (Sawchuk 1984; Sawchuk & Rector 1980) . The major pathophysiological functions which are affected and their pharmacokinetic implications are summarised in table 1. The rapid changes of clinical status can be associated with variations in the disposition of drugs. In the case of drug protein binding, it is clear that the same variable may be affected differently according to the type of drug (acidic or basic), and the degree of control of the clinical condition in the various stages. Whereas a marked decrease in the serum concentration of albumin occurs, possibly leading to an increase in the free fraction of acidic drugs, the concentration of ai-acid glycoproteins is increased, leading to increased binding of basic drugs (with consequent altered volume of distribution, serum concentrations and pharmacological effects [see Bowdle et al. (1980) for phenytoin, and Liebel et al. (1981) for d-tubocurarine]. However, these changes may be rapidly followed by normalisation of the situation when fluid and protein losses are compensated. Along this line of reasoning, a comprehensive approach has also been proposed for cardiovascular emergencies (see review by . This review is very exhaustive and should be referred to both for its methodology and the specific information it provides.

The general criteria and recommendations for fluid replacement therapy have been adequately reviewed (Shoemaker 1982) and will not be discussed here in detail. Precise measurement of fluid balance by weight is often clinically impracticable and must be based on information available from haemodynamic assessments, haematocrit values (also influenced by bleeding as well as fluid loss), haematocrit/Na+ ratio, plasma and urine osmolarity, and the urinary Na+/K+ ratio.

Considerable controversy still exists regarding the optimum fluid replacement schedule in shocked patients when adult respiratory distress syndrome (ARDS) is either feared as a complication or has already developed. The inconsistency of the available data could possibly be a consequence of the uneven quality and comparability of the clinical material represented in most publications (table I1) , where different therapeutic end-points may have been aimed at (e.g. normalisation of diuresis, systemic arterial pressure, central venous pressure or pulmonary artery wedge pressure). The following discussion is advanced as a possible basis for a consensus on fluid replacement policy.

Risk forARDS I. Young patients who are in a satisfactory general condition but hypovolaemic, or in a very early phase of shock: volume replacement therapy can be equally efficaciously and safely based on either crystalloids or colloids; however, a slower haemodynamic normalisation may be expected and a higher water and salt input is required with crystalloids.

2. Older patients in a similar clinical condition, presenting with cardiac, renal or pulmonary prob- ) Dahn et al. (1979 ) Johnson et al. (1979 Clinical condition Polytrauma Colloids equally effective as crystalloids Virgilio et al. (1979) Aortic surgery lower et al. (1979) Crystalloids harmful Jelenko et al. (1979 ) Boutros et al. (1979 ) Haupt & Rackow (1982 3. Patients in whom sepsis is the underlying cause of hypovolaemia, and young patients with advanced or very severe traumatic shock: colloids are again the first choice. In these conditions an increased permeability of the arteriolar-capillary (mainly pulmonary) membrane can be expected within 24 to 48 hours of the acute event. As particular attention must be paid to haemodynamic colloidal-osmotic equilibria, volume replacement should be assured with the smallest possible volume of crystalloids and albumin. The colloid osmotic pressure (COP) should never be lower than 17mm Hg (Morussette et at. 1979 ) and the colloid osmotic pressure-pulmonary artery wedge pressure difference must not fall below 6mm Hg .

In general, colloids (dextran or hetastarch) that are similar to albumin in molecular weight and in their preferential distribution in the intravascular compartment (Dawidson et al. 1980; Grundmann & Meyer 1982; Haupt & Rackow 1982) are the preferred choice. When artificial colloids are used, careful monitoring of the colloid osmotic pressure is essential to avoid the risk ofhypervolaemia leading to interstitial oedema. Neither for albumin nor for artificial colloids has the water and salt retention reported by Lucas et al. (1980) been confirmed (Haupt & Rackow 1982) .

The criteria for fluid replacement therapy in patients with acute respiratory distress syndrome are even more controversial. According to some authors (Lucas et al. 1980 ), colloids increase leakage from the interstitial space, where albumin binds to collagen, and impair lymphatic drainage. As a result, the interstitial colloid osmotic pressure rises, and the oedema of the alveolar-capillary membrane is worsened. An opposite view is centred around the role of albumin in sustaining the increase of intravascular colloid osmotic pressure: because permeability increases as a gradual and not generalised process, the interstitial fluid is 'dragged' in and allows re-expansion of the circulating volume, with a consequent lower requirement for water and salt loading, and a reduction of the risk of interstitial oedema (Pontoppidan et al. 1972; Wilson & Sibbald 1976) .

While a reliable direct measure of the extent of the leakage is difficult to achieve, comparative trials of crystalloids versus colloids in ARDS patients support the second hypothesis showing that colloids definitely produce more satisfactory haemodynamic stabilisation with the smallest water and salt load Hauser et al. 1980; Skillman et al. 1970 ). On the other hand, worsening of pulmonary function following excessive fluid loading is a well-known problem in severe ARDS . Colloids (whenever possible, albumin) appear to be the fluid of choice for haemodynamic stabilisation in 'extreme ARDS' (Gattinoni et al. 1980 (Gattinoni et al. , 1983 Iapichino et al. 1983 ).

The justification for this section may be summarised as follows (Askanazi et al. 1980a; Baker et al. 1982; Baracas et al. 1983; Birkhahn et al. 1980; Clowes et al. 1980a Clowes et al. , 1983 Kien et al. 1978; Long et al. 1977a,b; McMenamy et al. 1981a; Mullen et al. 1979; Stein et al. 1977 ): 1. There is a close relationship between the severity of acute injury (of whatever origin), energy 667 demand and the dynamic balance of protein synthesis and catabolism (nitrogen balance). 2. An uncorrected depletion of amino acids from the muscular, pulmonary and immunological reservoirs is likely to be associated with the development ofimmunodepressed/infective states and therefore with an unfavourable overall clinical course of the critically ill patient. 3. An adequate anabolic response (positive or less negative nitrogen balance) through nutritional support is a prerequisite for a favourable outcome . Table III summarises the experimental and clinical evidence for the role of total parenteral nutrition in assuring a favourable nitrogen balance by minimisation of the loss of proteins which are critical to assure essential physiological functions, and by control of catabolism and stimulation of anabolism. The present state of knowledge about the reciprocal role of calories and nitrogen (Bozzetti 1976; Elwyn et al. 1979; Iapichino et al. 1985; Jeejeebhoy 1977; Shizgal & Forse 1980) can be summarised as follows: a) At every caloric intake, the nitrogen balance improves with nitrogen supplementation b) Conversely, at every nitrogen intake, increased caloric support favours a better nitrogen balance c) Glucose or mixed glucose and lipid sources of calories can be considered equivalent in favouring nitrogen utilisation in depleted patients, at least after an adaptation phase (Bark et al. 1976; Jeejeebhoy 1977; Macfie et al. 1981; Shizgal & Forse 1981) . However, in the acute posttraumatic phase, a mixed glucose and lipid intake is associated with a worse nitrogen balance than is the case with an equivalent glucose intake (Freund et al. 1980; Long et al. 1977a; Shizgal & Forse 1981; Woolf son et al. 1979) . Calorie and nitrogen requirements differ according to both the needs of the individual patient and the target nitrogen balance. For example, in depleted non-catabolic patients, the main therapeutic goal is building of the lean mass (nitrogen balance +2 to +4 g/day): this will require 0.15 to 0.25g nitrogen/kg plus 40 to 60 non-protein kcal/ kg of actual bodyweight. On the other hand, injured patients require primarily control of the major catabolic nitrogen loss: in this situation a double amount of nitrogen (0.27 to 0.29 gjkg) plus 35 to 40 non-protein kcal/kg will favour a zero nitrogen balance (1 IU of insulin added to each 4 to 8g fraction of glucose assures a better anticatabolic effect) [Iapichino et al. 1982; Woolf son et al. 1979 ]. The achievement of a positive nitrogen balance in the acute reaction phase is an unrealistic target (see below). The role of specific aminoacids in injured and septic patients is still being investigated. Branchedchain aminoacids are essential to assure adequate synthesis of proteins in the liver (Blackburn et al. 1979; McMenamy et al. 1981 b) ; their hormone-like action in controlling the rate of muscular catabolism is still controversial (Cerra et al. 1982; Freund et al. 1982; Schmitz et al. 1982 ). On the other hand, a reduced supply of aromatic and sulphurated aminoacids is recommended since their utilisation in the liver is impaired in septic conditions (Clowes et al. 1980a; Freund et al. 1979; Larson et al. 1982; McMenamy et al. 1981 b; Smith et al. 1982 ).

A comprehensive definition of this condition would include all patients with impaired pulmonary gas exchange and partial carbon dioxide retention. The higher cardiovascular and respiratory demand may worsen a situation of acute respiratory insufficiency in spontaneously breathing patients with limited ventilatory capacity. Partial replacement of the glycidic load with lipids has been proposed to reduce the surplus production of carbon dioxide which follows the increased respiratory quotient (Askanazi et al. 1981) . The key feature is an increased energy requirement stimulated by the caloric intravenous supply (Askanazi et al. I 980b; Gattinoni et al. 1974) . This concept appears applicable in depleted patients with respiratory insufficiency where the protein-sparing effects of lipids are equivalent to those of carbohydrates; however, it is debatable in injured patients in whom lipids have been shown to have a lower proteinsparing effect when compared with glucose alone. Hence, to achieve the same nitrogen balance, a higher caloric intake will be required when using lipids (Jarnberg et al. 1981) .

In summary, the present state of knowledge underlines the importance of avoiding caloric loads higher than those strictly required for a zero nitrogen balance (a positive nitrogen balance must be discouraged) [Iapichino et al. 1983 ]. Optimisation of the nitrogen balance should be sought through increased nitrogen intake, provided due attention is given to the increased metabolic demands (Askanazi et al. 1982 ).

This clinical condition which follows shock, trauma or sepsis is characterised by an altered fluid, electrolyte and acid-base status, as well as by hypercatabolic activity which can lead to an impaired nutritional condition and high morbidity and mortality.

The suggested treatment of the nutritional deficiency is based on a nitrogen intake tailored to match the losses. Any type of aminoacids coupled 669 with at least 35 to 40 kcal/kg may be used: unlike the situation in chronic renal insufficiency, urea recirculation is not clinically relevant (Lee 1980) . This approach modifies the classic Giordano total parenteral nutrition schedule for acute renal failure. It must be noted, however, that the few randomised studies which have addressed this issue (Kopple & Feinstein 1983) have not confirmed the superiority of a free nitrogen input (15g nitrogen composed of equal parts of essential and nonessential aminoacids) over the 2 to 3g nitrogen essential aminoacids only regimen, in improving nitrogen balance and survival. No definite criteria can therefore be set for the choice and clinical use of branched-chain amino-or ketoacids to assure optimum catabolic control.

The caloric component can be assured with glucose or glucose-lipid mixtures, provided no more than I g/kg oflipids is given at a slow infusion rate to allow for their reduced elimination (Druml et al. 1982 ).

In acute, toxic and infectious liver conditions, the aim of nutritional support is dual (Fischer 1981 ) : I. To reduce the release from muscles of aminoacids the liver cannot metabolise (glucose and insulin are effective); and 2. To re-establish the plasma balance of branchedchain and aromatic aminoacids by infusion of branched-chain aminoacids. In chronic, severe hepatic insufficiency, nutritional therapy can playa more important role. In patients with hepatic encephalopathy, it is currently recommended (Fischer 1981 ) that a 24-to 36-hour infusion of high-dose branched-chain aminoacids be given to waken the patient, followed by supportive nutritional therapy based on glucose and aminoacids (60 to 70 g/day, mainly branchedchain aminoacids to avoid worsening or precipitation of encephalopathy which may occur with aromatic and sulphurated components). The awakening effect of branched-chain aminoacids in this situation is supported by data obtained from ex-perimental models (Higashi et al. 1981; Smith et al. 1978 ) and from uncontrolled clinical trials (Fischer et al. 1974 (Fischer et al. , 1976 ; however, the results from controlled clinical trials are more doubtful (James et al. 1979; Michel et al. 1980; Wahren et al. 1981) . Furthermore, the overall efficacy of the treatment seems to be restricted to an improvement of the general clinical status. No long-lasting benefit can be claimed with respect to brain function and survival (Eriksson & Wahren I 982).

The aims and terms of reference for the metabolic aspects of enteral nutrition are the same as for total parenteral nutrition (with which enteral nutrition is often combined) and will not be discussed here. Enteral nutrition has acquired an increasing role in almost all severe (Kaminski 1976; Luc et al. 1981 ) and even extreme (lapichino et al. 1983 ) clinical conditions, for 2 reasons: 1. The development of techniques, devices and products which ensure adequate and easily manageable nutritional intake; and 2. A better understanding of the physiological and biochemical mechanisms underlying enteral absorption of the various components of a diet. Given as an oral meal, enteral nutrition is an integration of a free diet which cannot assure more than 80% of the body's caloric and protein needs. Given through a nasogastric tube, it may represent the only source of nutntional support or an effective complement to total parenteral nutrition. This latter combination reduces both the water load of enteral nutrition, and total parenteral nutritionrelated complications, and is specifically useful for oliguric, cardiac, ARDS and head injury patients.

Adequate oxygenation is often a critical issue for the intensive care clinician. There is no doubt that low arterial oxygen levels are extremely dangerous in acute situations: 'hypoxia not only stops the machine, it also destroys the machinery. ' Ventilatory management is still the cornerstone 670 of therapeutic intervention in intensive care patients. The provision of a viable gas exchange represents the immediate therapeutic target, but undoubtedly healing of lung lesions and prevention of damaging effects upon other organs remain the ultimate goals and the ones which should always receive maximum attention. ARDS can be taken as the model of a relevant clinical condition where the problems of respiratory care can be discussed. In ARDS lungs, the normal matching of ventilation and perfusion is greatly altered. Under these conditions, the necessary adequate compromise between oxygenation and CO2 clearance may require extremely unphysiological interventions, often demanding high minute volume ventilation (sometimes in excess of 20 L/min), high airways pressures, and high inspired oxygen fractions (Fj0 2). As there is often no specific treatment available for ARDS, management is confined merely to life-supportive intervention (Gattinoni et al. 1983 ).

Mechanical ventilation was introduced to support the breathing of patients with neuromuscular disease and those undergoing paralysis anaesthesia. Its use soon spread to the treatment of what was later to be called ARDS.

It must be emphasised that the mechanical ventilator is not primarily an oxygenator, but rather a mechanical pump that removes CO2 from the natural lung. However, oxygenation does not require ventilation, as anaesthesiologists have long shown with the apnoeic oxygenation technique (Frumin et al. 1959) . Table IV provides an overview of the pros and cons to be expected from some of the available ventilatory management measures, while figure I summarises the pathophysiology involved in the main aspects of respiratory support, focusing on improvement of oxygenation. It is important to note that the only relevant benefit brought about by controlled mechanical ventilation (CMV) is the relief of respiratory work, as long as it does not induce muscular atrophy and discoordination.

The role of CMV is diminishing as ventilatory (Marshall et al. 1982) PEEP (Shapiro et al. 1983; Weisman et al. 1982) Increased Paw (Boros et al. 1977) Increase alveolar recruitment Increase lung volume Improve oxygenation Increased VO/VT Increased barotrauma Decreased cardiac output and renal function (Marshall et al. 1982) Increased Fj0 2

Correct local alveolar hypoxia in low VA/Q areas Oxygen toxicity (Deneke & Fanburg 1982) Atelectasis (denitrogenation) [Dantzker et al. 1975 ) Increase the oxygen content of blood management is becoming more a combination of specific therapeutic interventions, with preset physiological goals. Thus, CMV has become a specific measure for patients in whom CO2 clearance cannot be safely reached by other means. It is often possible to combine two or more manoeuvres to reach a specific goal, e.g. positive end-expiratory pressure (PEEP) plus CMV is called controlled positive pressure ventilation (CPPV), offering the advantages (and disadvantages) of the two.

As is often the case in intensive care medicine, the therapeutic manoeuvres indicated for ARDS carry a high rate of side effects. At present, the iatrogenic effects of therapy upon the course of the lung disease cannot be clearly separated from the natural history of ARDS (Pratt et al. 1979 ). On the other hand, respiratory therapy is a mandatory, lifesupporting measure without which ARDS can progress to a full-blown situation. Hence, even when trying to abstain from an Fi0 2 higher than 0.4 to 0.6 and from higher than normal airways pressures (20cm H20 PEEP), we are sometimes left with no other choice than to resort to more dangerous set-tings to ensure viable blood gases. The frustrating exercise of balancing advantages and disadvantages of any specific intervention should always be performed. Unfortunately, little is known about the long term effects of respiratory treatment upon other systems. It is not uncommon for ARDS patients to die not from hypoxaemia, but from failure of other organs or systems (Kirby et al. 1975 ) whose function may have been severely affected by the respiratory therapy.

Acute cerebral damage due to focal traumatic, non-surgical lesions has been chosen as the reference clinical condition for a discussion of the basic principles (pathophysiological mechanisms and criteria for pharmacological treatment) involved in formulating a rational clinical pharmacological ap- proach to intensive care situations with major involvement of the central nervous system (CNS). However, the points made here have a broader application to cerebral lesions frequently occurring in intensive care settings that are caused by other noxae such as focal vascular injury, acute infections and tumours. The basic processes leading to acute cerebral damage can be described by the scheme shown in figure 2. The hyperaemic swelling and ischaemic focal cerebral oedema cause a high intracranial pressure (fCP). Focal oedema leads to cerebral shift; the displacement of the brain from its axis against the tentorium and the falx leads to cerebral ischaemia and oedema. The resulting diffuse oedema further increases the intracranial pressure and the ischaemic anoxic damage (Bruce et at. 1981; Clifton et al. 1983; Cold & Jensen 1978; Miller 1985; Obrist et at. 1984; Overgaard & Tweed 1983) .

The variables listed in figure 2 may be amenable to prophylactic and therapeutic treatment according to the following sequence, which serves as a guide for discussion of the most promising, albeit controversial, approaches to the management of acute cerebral damage: I. Prophylaxis and treatment of the causes of secondary cerebral damage (worsening factors) 2. Avoiding or minimising cerebral oedema and ischaemic anoxic damage 3. Symptomatic therapy of high intracranial pres-sure if the above 2 measures fail Prophylaxis and Treatment of Worsening Factors 673 4. Preservation of cardiovascular, respiratory and metabolic haemostasis 5. A voiding iatrogenic problems of late complications 6. Early rehabilitative procedures.

The primary goal of the management of acute cerebral damage is to control all factors which could worsen the cerebral lesion. Table V The hyperdynamic/hyperadrenergic syndrome (table VI) may be seen as an exaggeration or caricature of the classical 'fight or flight' reaction; it is also seen in the narcotic 'withdrawal syndrome'. A hyperadrenergic state is the common feature of these conditions. In patients with acute cerebral damage this may be due to the functional disconnection of the brainstem from the hemispheres or to a direct stimulation of the diencephalic-hypothalamic-brainstem system as a result of high intracranial pressure, ischaemia, blood leaking into the CSF, pH changes in the CSF, etc. (Pia 1974) , as well as extracranial causes. Therapy and prophylaxis of this syndrome rest upon the rather unspecific use of various classes of sedatives and of antiadrenergic drugs. Although their mechanisms of action are different and the treatment schedules are far from well established, all the drugs listed in table V have an unspecific 'sedative' effect, permit good control of the increased muscular tonus and of decerebrate-decorticate fits, and contribute to the reversal of the systemic and cerebral circulatory and metabolic alterations.

Seizures: Convulsions are frequent and very dangerous in acute cerebral damage. Effective prophylaxis is not easy because of the short interval between injury and the first post-traumatic seizures. Standard treatment is based on the classical anticonvulsant drugs, as shown in table V. Though almost universally accepted, the efficacy of prophylaxis has never been proven in formal trials ). The dosage schedules pose an interesting clinical pharmacological problem, as early 'effective' drug concentrations do not seem to be easily defined or achievable, even when high loading doses are used. Moreover, it is clear that the prophylactic benefit could be the result of the various measures included in the overall management scheme rather than of a specific pharmacological prophylaxis. Abbreviations: CBF = cerebral blood flow; CBV = cerebral blood volume; ICP = intracranial pressure; CMR02 = cerebral metabolic rate for oxygen; VA/Q = ventilation to perfusion ratio,

Available knowledge on commonly adopted treatment for cerebral oedema is rather inconsistent. The standard regimen of corticosteroids for 3 to 4 days is based on theoretical considerations and suggestive experimental data, but has never been confirmed in properly controlled clinical trials (Braughler & Hall 1985) [table VII]. Positive results in other clinical conditions such as peritumoural and encephalitic oedema are hardly transferable to traumatic acute cerebral damage. A cornerstone of the treatment is the control of hyperaemia, a key causal factor in oedema. While there is at present no effective method for influencing the 'luxury flow' locally, antiadrenergic drugs (see table V) can help by reversing the hyperdynamic state (Clifton et al. 1983 ). This treatment requires careful circulatory, respiratory and EEG monitoring but compares favourably in terms of unwanted effects and nursing requirements with other treatments aiming at the same goal, such as barbiturate-induced coma or protracted mechanical ventilation (see below).

Control of High Intracranial Pressure From figure 2, it is clear that the intracranial pressure depends on the type and the severity of the various factors which constitute the clinical condition of acute cerebral damage (Shapiro 1975) . However, intracranial pressure monitoring (itself not a completely safe or reliable invasive technique) has failed to assess the relationship, if any, between intracranial pressure, the severity of cerebral damage, computerised tomography scan images, and the outcome. The threshold itself for treatment is not well defined.

Nevertheless, immediate and aggressive therapy is indicated on clinical grounds in the presence of cerebral shift and when cerebral coning is impending. The intervention strategy is summarised in table VIII (Quandt & de los Reyes 1984) . Osmotic diuretics are the treatment of choice in the acute phase to reduce rapidly the extravascular/water by increasing plasma osmolality. However, their prolonged use is best avoided; osmotic diuretics accumulate in the brain across a damaged blood/brain barrier, reversing the osmotic gradient, with a potential oedema-promoting effect. They also disturb the fluid and electrolyte balance, which must be the real target of prolonged and meticulous control. Frusemide (furosemide) may be useful in order to prevent the initial temporary hypervolaemia caused by osmotic diuretics and to reduce CSF formation.

Mechanical hyperventilation can reduce intracranial pressure in a few seconds by hypocapnic vasoconstriction and, in part, by decreasing cardiac output via positive intrathoracic pressures. Both mechanisms quickly fade, the former because of the restoration of normal cerebral pH, the latter because of compensatory water and salt retention (Heffner & Sahn 1983) . Barbiturate-induced coma has been proven to reduce high intracranial pressure in some patients unresponsive to mechanical hyperventilation, osmotics, corticosteroids and CSF drainage (Marshall et a1. 1979a,b; Rockoff et a1. 1979) . Interestingly, the pharmacokinetic behaviour of the most frequently studied drug, pentobarbitone, has only very recently become the object of specific interest (Bayliff et aI. 1985) . Repeated barbiturate doses are associated both with diagnostic problems and various systemic complications (e.g. hypotension, infections, bedsores). As severe cardiovascular depression is possible, the use of barbiturates calls for simultaneous monitoring of intracranial pressure and blood pressure in order to maintain a safe cerebral perfusion pressure. Since the reduction of intracranial pressure produced by these agents is related to a decrease of the cerebral metabolic rate for oxygen (CMR02) and a consequent decrease of cerebral blood flow (CBF) and cerebral blood volume (CBV), no further effect can be expected when metabolism is severely depressed and a 'burst suppression' pattern on the EEG is achieved (Kassell et al. 1980) . Moreover, only the evoked potentials are useful when testing the neurological state in barbiturate-induced coma. When cerebral compliance is critical, bolus doses of sedatives can be of benefit to prevent the increase of intracranial pressure in response to stimulating procedures (Moss et a1. 1983) ; in these cases, EEG monitoring can be useful to predict their effect on intracranial pressure and cerebral perfusion pressure (Bingham et a1. 1985) . The persistence of a high intracranial pressure, in the absence of surgically treatable lesions, means that treatment has failed, either due to inadequacy or delay. received much experimental and clinical attention over the last few years, has been associated with good results in focal ischaemic anoxic lesions but not with respect to global cerebral damage (Safar 1980) . The overall rationale of an intervention which assigns beneficial effects to a reduction of the metabolic rate is under revision, as the functional metabolic rate is often already depressed in these patients (Astrup 1982) . Table VIII summarises the goals (and the state of knowledge about the means to achieve them) of this strategy, which can be defined as 'a safe means of increasing the brain's tolerance to the anoxicischaemic insult' (Cohen (981) .

As other experimental suggestions have failed in clinical trials, prevention and treatment of brain ischaemia remains a frustrating and unresolved clinical problem (Hinds 1985) .

Despite the relatively long series of attempts centred on this procedure, it is still largely idealistic. Barbiturate-induced coma therapy, which has Over the last few years, new therapeutic strategies have been based on suggestions arising mainly from the availability of new technological tools which may provide a better insight into the morphological, metabolic and vascular aspects of acute cerebral damage. Relatively little has been done to exploit a parallel body of knowledge on neurotransmitters, which has grown significantly, but which up until now has resulted only in rather nonspecific applications. Table IX summarises current knowledge of the effects of two recently emergent drugs in this area, clonidine and naloxone. The interactions of suggestive clinical evidence, pharmacological and biochemical background, and pathophysiological hypotheses indicate the probable path of clinical pharmacology over the next few years in this field. of fatalities after the first week of intensive care unit stay are associated with infectious complications) [Allgower et al. 1980; Machiedo et al. 1981; Pottecher et al. 1979] . It is also well known that anti-infective (mainly antimicrobial) therapy accounts for the greater part of the most frequently prescribed drugs (Buchanan & Cane 1978; Farina et al. 1981) .

While the importance of appropriate antimicrobial treatment in improving survival in many critical care conditions is undisputed, an analysis of the studies in this field (table X) suggests a situation where non-pharmacological factors are in the forefront. Where mentioned, antibiotic prophylaxis or treatment appears mostly as an accessory in descriptive and controversial analyses of hard-to-compare measures applied in widely differing settings. Sufficient prospective, properly stratified data (according to the various clinical and environmental variables) are not available to support a well-defined strategy and, even less, specific drug treatment. It is interesting that one of the most optimistic reports which documents the critical role

It is a commonly accepted view that infections play a major role in the morbidity and mortality profile of intensive care patients (more than 50% of the setting in improving the infection rate and severity has little to say for antibiotics (Goldman et al. 1981 ). On the other hand, the two more drugoriented studies (Pottecher et al. 1979; Stevens et al. 1974 ) which failed to provide evidence of a positive role of antibiotics in improving survival and decreasing the spread of infections must be criticised for their poor clinical pharmacological approach, which could have been one of the reasons for their 'negative' results. Surgical drainage and eradication of the focus is obviously the single most important and effective step in the treatment of infection (Meakins et al. 1980; Rapin & George 1983) . Hence, the major challenge for antimicrobial treatment comes from infections such as pneumonia, bacteraemia or meningitis which are not amenable to surgical intervention. The state of the art in this field can be summarised in 6 points, which also provide a possible framework for the urgent, but difficult task of producing reliable data.

1. The host-environment relationship where a pharmacological anti-infective measure is taken is the most decisive factor in determining the overall outcome. Table XI provides a problem-oriented

Infection control programme, nurse epidemiologists and subrapubic bladder drainage improved infection rate Staffing and environment play critical role Research needed to explore cellular mechanisms likely to be the key of sepsis-related organ failure Doubtful role of antibiotics, which should be used as late as possible Priority given to early and aggressive treatment of respiratory and circulatory failure Critical role of definitive surgery, immunological defence, adequate nutritional support Various systemic antibiotic regimens did not increase survival in acquired pneumonia; ?polymyxin aerosol in prophylaxis Survival improved after proper antibiotic treatment (retrospective study) Doubtful efficacy even of appropriate pharmacological intervention Antibiotic drug efficacy in about 45% of cases; mainly minor complications guide to the major risk factors, which must be considered critical variables when instituting and evaluating prophylactic and/or therapeutic treatment.

2. The pathogenetic potential of an intensive care unit environment -where invasive procedures and the risk of cross-infection are associated with the high vulnerability of the critically ill patient -should be combated with careful preventive measures. The well-documented, mandatory preventive measures shown in table XII are largely the same as the various recommendations based on clinical common sense, and do not provide a ready-to-use solution (Eickhoff 1981) .

3. Clinical pharmacology is often held up as important in setting optimum guidelines for therapeutic schemes. Such confidence is based on an expected ability to describe and predict the pharmacokinetic behaviour of various drugs which are handled in the body by organs whose function is often rapidly changing, and which have to find their way into tissues and organ 'sanctuaries' where the infecting organisms exert their effect and multiply. Standard pharmacokinetic guidelines and monitoring of serum concentrations (and in selected cases, of other tissues or biological fluids) must be applied in cases of organ failure (Bennet et al. 1977) . The expectation is fully justified, but for the time being it still needs confirmation in routine clinical settings. 4. Clinical microbiology may playa more important role than clinical pharmacology. Virtually all micro-organisms have been found to be potential pathogens in the critically ill and any tentative ranking of their importance may be misleading. The clinical condition, the setting, and the problems of bacterial selection, emergence of resistant strains, transfer of resistance and transfer. of resistant strains require comprehensive, attentive surveillance to detect the agent responsible in good time. Clinical microbiology specific to the intensive care situation is faced with apparently insurmountable obstacles, judging from the scant available information. Clinical experience tends to bear out this unsatisfactory state of affairs, because of the objective difficulties in distinguishing harmless colonisation from harmful infection, the incompatibility between the urgency of clinical questions and the time microbiologists need to find relevant answers, and the sophistication required to sample 681 and grow representative and reliable specimenswhich is complicated even further by the presence of mixed infections. 5. Prophylaxis, the classical controversy of antibiotic usage (short of hard specific data for most intensive care conditions), must be based on the principles accepted in other areas of medicine: short term perievent treatment with the narrowest spectrum agents should be the rule. Only experimental data suggest the usefulness of prophylaxis even some hours after contamination (Miller & North 1981) .

6. Life-threatening conditions, where very active and timely anti-infective treatment could be crucial, have to be faced, de facto, on mainly empirical grounds. The recommendations in table XIII summarise current knowledge with respect to how a specific condition of the host can most appropriately be treated. Amikacin or 'third generation' cephalosporin or azlo-, mezloor piperacillin (± vancomycin) Antimycotic treatment a Treatment is not adequate for infections due to mycobacteria, fungi, Mycoplasma, or Legionella sp. and similar causal agents of interstitial pneumonia. Resistant strains must be considered (e.g. Pseudomonas sp., S. aureus). b Treatment is not adequate or optimal for infections due to Gram-positive aerobes, some anaerobes, mycobacteria, Pneumocystis carinii, or Mycoplasma.

The three main variables which interplay in determining the dynamic, partially self-regulating, circulatory status are: (I) the circulating volume;

(2) cardiac function; and (3) extension of the vascular bed, which in intensive care patients may be specifically affected by any of the most frequently occurring critical conditions listed in table XIV. The type and the priority of the various forms of pharmacological and non-pharmacological treatments should be evaluated against this background. A few general principles are worth emphasising before focusing on specific measures: 1. Complex and multiple interactions are the rule in cardiovascular emergencies arising in the intensive care unit. The rationality of the therapeutic approach depends directly on the degree of accuracy achieved in ranking and correlating the variables to be taken into account. 682 2. Rapid, spontaneous modifications of the haemodynamic status call for particular caution in passing a positive or negative judgement on anyone pharmacological treatment. A conservative attitude is mandatory when no carefully controlled comparative data are available to support the role of new drugs or treatments. 3. An intelligent surveillance strategy may be more useful and informative than active intervention. Intrinsic compensatory circulatory adjustments, though abnormal (e.g. tachycardia in hypovolaemia or hyperthermia) may be satisfactory and in fact better tolerated in certain clinical situations than prompt 'normalisation'. 4. Adequate tissue perfusion (when organ function is threatened) and circulatory stability (when compensatory mechanisms are at risk of breakdown) are the key terms of reference for any intervention strategy. The most common situations requiring prompt ad- justment are those involving imbalance between the circulating volume and the size of the vascular bed, with impending or actual circulatory failure.

Adequate fluid replacement is of prime importance in all cases in which intravascular volume is reduced. It is important that fluid replacement be 'adequate', not merely 'normal', in order to correct a pathological circulatory pattern. For example, in a situation of generalised loss of vasomotor tone (see table XIV) fluid replacement must be higher than normal, until physiological or pharmacological measures reduce the abnormal extension of the vascular bed.

Vasoactive drugs are important adjuvants to fluid therapy, as they help to restore and maintain a stable haemodynamic setting by modifying the extension of the vascular bed. Their rational use depends on the accuracy of information available on circulatory patterns and on timely monitoring of the patient's clinical conditions. The main intensive care conditions where these drugs are used are: I. During volume replacement, to maintain adequate perfusion pressure to vital organs as long as necessary 2. In patients with poor cardiopulmonary function, where rapid volume loading could be harmful 3. In prolonged depression of vascular tone with abnormal expansion of the vascular bed requiring sustained fluid replacement to prevent excessive loss of water into the extravascular space 4. In some stages of septic shock with very low peripheral vascular resistances, to restore the normal extension of the vascular bed 5. For dopamine alone, at very low doses (dopaminergic effect) to improve renal blood flow in renal insufficiency due to circulatory failure.

Primary pump failure is a very rare event in patients admitted to a general intensive care unit. Most often, multiple factors lead to secondary depression of cardiovascular function (see table  XIV ). In most instances cardiac failure is seen as the final stage of multisystem organ failure.

A rational therapeutic approach must be to remove, or treat, the aetiological factors; a vasoactive drug with inotropic properties should be considered only as a temporary support for the vascular system until it is completely stabilised. Inotropic agents are often administered in clinical conditions (e.g. the postoperative period, pulmonary embolism, respiratory insufficiency, mechanical ventilation with PEEP) which do not represent specific indications, and where the efficacy of drug treatment has not been confirmed in properly conducted clinical trials (see also below).

Digitalis is best avoided (Herbert & Tinker 1980 ) because of its weak inotropic action and because in the many abnormal pathophysiological situations encountered in patients in the intensive care unit, its pharmacological action may be altered, with a resultant increased incidence of toxic effects (Opie 1980) . In rare conditions when myocardial contractility is selectively depressed (e.g. in drug intoxication; see table XIV), dobutamine, glucagon or isoprenaline (isoproterenol) are preferred, provided no associated severe arrhythmia is responsible for the vascular failure. Table XV shows the most common conditions which can be associated with cardiac rhythm disturbances in general intensive care patients. With the obvious exception of life-threatening arrhythmias, general supportive measures of intensive care, without specific drug treatment, provide a satisfactory solution. Antiarrhythmic therapy is a matter of clinical judgement, and should be based on aetiological factors, the background of a particular illness, and the patient's haemodynamic balance. To avoid inappropriate or unjustified use of drugs, priority should always be given to removing all possible causes of arrhythmias. This is best illustrated with three examples: I. Arrhythmias due mainly to abnormalities in the ventilatory pattern: normalisation of ventilation 

The basic pathophysiological condition in pulmonary embolism is a mechanical impairment of the right ventricle outflow, with pressure overload on the same chamber. The dilated right ventricle reduces left ventricular compliance and the diastolic filling volume (Mcintyre & Sasahara 1974) . The more severe the pulmonary obstruction, the less is the left ventricle inflow, which leads to a low cardiac output and shock. Because cardiac contractility is not impaired, there is no rationale for inotropic drugs per se. but if shock is present, vasoactive drugs should be considered to sustain blood pressure and to assure blood flow to vital organs. Efforts should be focused on reducing the degree of obstruction in the pulmonary circulation.

Adequate tissue perfusion and duration of shock are the main factors which influence survival (Weil & Nishijima 1978) . The immediate replacement of lost fluid therefore has priority. Low cardiac output with adequate filling pressure is usually observed in the late stages of multisystem organ failure. Although inotropic agents are generally used, their long term efficacy is not proven and the prognosis depends mainly on whether prolonged damage to vital organs has occurred.

In recent years, many studies have demonstrated the pathological consequences of mechanical ventilation with PEEP on the circulation. In ARDS, both PEEP and right ventricular stress markedly alter the interventricular relationship, leading in the most severe cases to 'left ventricle tamponade' and a fall in cardiac output (Laver et al. 1979) . As in major pulmonary embolism, the low output is due to mechanical abnormalities and does not itself require inotropic support. However, vasopressive agents ate widely used to sustain blood pressure and counteract the negative haemodynamic effect of PEEP (Hemmer & Suter 1979) .

The use of vasodilators (such as sodium nitroprusside) to 'unload' the right ventricle (Petty & Fowler 1982) , though theoretically acceptable, can-not claim reliable data confirming its lasting efficacy. In the late stages of ARDS, severe anatomical damage in the pulmonary circulation dictates the poor prognosis, and vasoactive or inotropic agents are unlikely to improve the outcome (Pontoppidan & Rie 1982) .

The control of haemostasis processes, either in terms of prophylaxis or treatment with drugs or blood components, is in the forefront of the clinical concerns in the intensive care unit setting, mainly for two types of patients: (a) those admitted for specific and severe haemostatic (haemorrhagic or thrombotic) problems; and (b) those who present with pathologies at high risk of haemostatic complications.

Two key drug groups, anticoagulants and fi-Massive transfusion with stored blood • Loss and dilution of coagulation factors and platelets • Possible role of stored blood in aggravating DIC 685 brinolytic agents, are most often used. Antifibrinolytic drugs are very rarely indicated or needed, and will not be discussed here. Because they are still advocated and used, it is worth stressing that the so-called 'haemostatic drugs' should be left out of any therapeutic or prophylactic armamentarium, as they lack any theoretical or clinical foundation (Verstraete & Vermylen 1982) . A common problem lies in the use of blood components, whose main clinical application is in massive blood loss. This is a frequent event in intensive care practice which besides being lifethreatening per se may be associated or followed by haemostatic failure through any combination of the events or factors listed in figure 3.

The risk of a situation changing from one of massive blood loss to one of haemorrhage must be met: (a) through comprehensive clinical surveillance of sequence B in figure 3 (whose evolution is mainly dependent on the underlying pathology); and (b) through the rational use of a transfusion strategy.

The key elements of clinical management of massive blood loss are summarised in table XVI, which is also a reminder of the corresponding discussion in section 3.3 (with respect to circulating volume and cardiac output). Stored whole blood is still widely considered to be the only readily available form of red cells and volume (and for this reason it is shown alone in figure 3 ), the only caveat to this being that when storage is longer than 24 hours it contains non-functioning platelets, and no more than 10% off actor V or 20% off actor VIII (Benson & Isbister 1980) . Fresh frozen plasma contains the natural inhibitors of coagulation and fibrinolysis (antithrombin-III and antiplasmin) as well as the coagulation factors. It therefore has the capacity of reducing the loss and dilution of coagulation factors, and of preventing or correcting possible disseminated intravascular coagulation (DIC) with its sequelae of multiple organ failure.

The role of fresh blood is still controversial (Counts et at. 1979; Loong et at. 1981 ) with respect to its definition (blood stored for less than 24 hours, or warm blood which is not anticoagulated). The few cases who have responded only to blood obtained by direct transfusion, after all haemostatic factors have been corrected, suggest that unknown factors of haemostasis are transfused with fresh warm non-anticoagulated blood (Editorial 1976; Sheldon & Blaisdell 1975 ). Because of the inherent risk of any direct transfusion, this practice should be reserved for the very rare 'resistant' cases and evaluated with ad hoc research protocols.

Disseminated intravascular coagulation can be seen as an intermediary disease mechanism, which can occur in a variety of clinical conditions, where blood coagulation is triggered by one of the factors shown in figure 4. This results in fibrin deposition in the microcirculation with subsequent organ failure.

Haemorrhage may appear when the consumption ofphitelets and/or coagulation factors exceeds the rate of synthesis (Mant & King 1979; Preston 1982) . The rationale for any intervention strategy is to address simultaneously various goals which can be ranked as follows (Verstraete & Vermylen 1982) : 1. Assure adequate tissue perfusion and oxygenation 2. Remove the triggering cause, if known and accessible 3. Break the chain of events shown in figure 4 4. Replace deficient factors.

Guidelines for the first two goals (I and 2) are covered by standard principles of treatment of multiple organ failure, of ablative surgery (e.g. in case of septic abortion or tumour masses), and of general pharmacological intervention (e.g. for sepsis or eclampsia). For the purposes of this discussion, the latter two goals (3 and 4) merit specific attention as being those where specific pharmacological treatment may playa role. The scant information available can be summarised as follows: I. The efficacy of heparin in positively influencing the morbidity and mortality associated with dis- seminated intravascular coagulation is far from proved. A possible role is seen:

(a) in some cancer patients;

(b) in obstetric situations (once the intrauterine cause is removed); (c) as a third-line drug, when plasma plus antiplatelet drugs and plasmapheresis have failed and; (d) in thrombotic thrombocytopenic purpura and haemolytic uraemic syndrome. As a general rule, heparin should be reserved for severe haemorrhagic complications and/or vascular thrombosis, with or after the administration of blood components. 2. When replacement of deficient (consumed) factors is needed, fresh frozen plasma is the firstline treatment, as in more complex conditions than a simple lack of fibrinogen, it provides the natural inhibitors of coagulation and fibrinolysis, antithrombin-III and anti plasmin, as well as coagulation factors. It also serves as a plasma expander. 3. Platelet concentrates may be useful when the platelet count is ;:s; 30-50,OOO/"d and consumption has not stopped.

4. Data on antithrombin-III (which may block intravascular coagulation without affecting local haemostatic processes; Liebman et al. 1983 ) are presently too scanty, and restricted to patients with severe hepatic failure, to permit the formulation of precise guidelines for its use. Antifibrinolytic drugs are contraindicated, as it is irrational to try to stop a natural defence mechanism against vascular occlusion. Moreover, a definite risk of aggravation of the patient's condition may follow, such as the evolution of haematuria into anuria. Concentrates of blood clotting factors (e.g. fibrinogen, cryoprecipitates, prothrombin complex concentrates) should also be avoided since they may contain activated clotting factors and therefore may aggravate the progression of disseminated intravascular coagulation.

Pulmonary emboli result in partial or total occlusion of the vascular bed. The basic pathogenetic feature is a mechanical obstacle to pulmonary blood flow, causing a higher right ventricular workload. In the most severe cases, the decrease in pulmonary blood flow leads to diminished left ventricular filling, which may result in a low cardiac output and shock. Clinically, this translates into only a minority of patients (27%) surviving the first hour after the event (Bell & Simon 1982) and in whom management can be offered (as outlined in table XVII). Possibly also because of the difficulties inherent in a satisfactory standardisation of the diagnosis, the very few properly controlled trials that have been conducted have been in small and largely non-comparable populations (Ly et al. 1978; Tibbutt et al. 1974; UPET 1973) , and their results may be seen at best as a contribution to serendipidous guidelines:

1. Thrombolytic therapy [streptokinase, urokinase, or one of the newer agents such as tissuetype plasminogen activator (Collen 1983) which have been tested up until now mostly in coronary conditions] is the first choice when the haemodynamic impairment requires rapid removal of the obstacle from the pulmonary vascular bed. 2. Anticoagulant therapy has been included in clinical trials as a mandatory follow-up for maintenance of the thrombolytic effect and must be seen as a key component of management aiming at the prevention of further emboli.

3. Haemorrhagic side effects may follow both thrombolytic and anticoagulant therapy. The true incidence is unknown, as the rates reported in randomised clinical trials derive from unusually well controlled and selected patients. It is worth stressing that heparin (Porter & lick 1977) is no less, and possibly even more likely to cause haem orrhagic side effects than thrombolytic agents. The bleeding which may follow administration of thrombolytic drugs is mostly a direct result of the invasive procedures needed to follow the evolution of the thrombotic process and the haemodynamic setting.

4. Surgery for embolectomy is the exception (mortality is still unacceptably high) and must be considered only when cardiac arrest follows or is associated with massive pulmonary embolism, or in the presence of absolute contraindications to anticoagulant and thrombolytic drugs in haemodynamically compromised patients. Vena caval interruption should be reserved for patients in whom anticoagulation is absolutely contraindicated (or anticoagulant and/or thrombolytic therapy has failed) and a further embolus would be life-threatening (Bell 1982) .

Prevention of pulmonary embolism in severely traumatised patients is of specific interest since: I. Severely traumatised patients are at high risk of thrombotic complications (Coon 1977) . This is due to severe tissue damage and subsequent blood clotting activation, coupled with other risk factors (e.g. immobilisation, possible negative water balance, possible surgical intervention). 2. Severely traumatised patients may be or are at risk of haemorrhage. Even minor bleeding may be very serious when localised in particular regions (e.g. the brain, pericardium, spinal cord). 3. Practically no information has been obtained directly in this group of patients. 'One must decide on an approach to prophylaxis by extrapolation from studies of other conditions, a perilous undertaking' (Salzman & Davies 1980) . The last caveat is even more challenging, and worrying, when the currently used prophylactic treatments are considered (Salzman & Davies 1980) . Oral anticoagulants fare better than other treatments (though their benefit has not been proven in terms of mortality reduction, possibly because of the small size of the tested populations), but carry the highest risk of severe haemorrhagic complications. This calls for the closest monitoring of laboratory variables to avoid mainly too low, and therefore ineffective, dosages. These drugs are further seen as controversial in orthopaedic and trauma patients and cannot be used in such clinically important categories as patients with head, pericardial, retroperitoneal or medullar trauma.

A rationale for low dose heparin (whose value has been proven in cases of elective surgery) does not exist, since significant amounts of circulating activated blood clotting factors are present when the patient comes to medical or surgical attention (Blaisdell 1979) . Dextrans, which do not require laboratory monitoring, could be a useful choice, but 689 available data (referring to populations treated with two different molecular weight molecules) do not provide any consistent guidelines with respect to efficacy, dosage or duration of treatment (Bell 1982) . While the risk of allergic reactions can be definitely considered low, haemorrhagic complications and the risk of fluid overload could become clinically relevant. Other procedures including 'physical' manipulation (physiotherapy, muscular electrostimulation, etc.) have been tested in nontraumatised, very low-risk patients.

Clinical conditions encountered in general intensive care units represent a uniquely challenging situation for what should be considered a common goal of clinicians and clinical pharmacologists: the application of scientifically sound criteria to the routine care of patients and, at the same time, to the development of new knowledge on the many unanswered questions concerning overall management and the use of specific drugs and treatment strategies. Our collaborative analysis of the published literature and of the approaches adopted in the various settings has underlined the basis from which this review originated: there is an urgent need for studies which allow the evaluation of individual treatments in the broader context of overall care. The philosophy and the sequence which have been adopted for the presentation and discussion of the various critical care situations seem to offer a good basis for revising practices used in different centres, and for defining the problems to be considered when planning prospective or retrospective controlled therapeutic or prophylactic interventions.

The traditional core of clinical pharmacology, clinical pharmacokinetics, should be seen: (a) as a very simple series of descriptive data sets on the pharmacokinetic behaviour of single compounds, which should be made available to clinicians working with intensive care patients, mainly to avoid drug-related side effects and in certain specific cases (e.g. the use of barbiturates) to allow a better control of the treated problems; and (b) as a conceptual basis for understanding and investigating, through monitoring of the pharmacokinetic behaviour of drugs, the modification of physiological compartments and functions.

The interplay and the interdependence of the main background conditions which are found in critically ill patients (altered nutritional status, fluid balance and respiratory function) with specific organ-related problems appear to have been up until now scarcely investigated. In this respect, it is interesting to note that the clinical syndrome ARDS may be described and approached differently depending on the expertise of the clinician in charge of the affected patient (Editorial 1986) . In this review an attempt has been made, within the individual areas, to identify a logically common strategy where the order of priorities is stressed, and where drugs may be clearly appreciated as a dependent variable.

The particularly complex situations which are the rule in general intensive care settings have so far favoured studies based on single-centre protocols, where standardised conditions of observation and treatment are assumed to be more easily assured. The advantage of this approach is obvious with respect to standardisation. However, the adoption of multicentre protocol designs for experimental and controlled evaluations of the outcome of different routine management strategies seems to be a worthwhile alternative to allow for the assessment of therapeutic and prophylactic interventions on larger populations. Appropriate stratifications could then offer new opportunities for understanding the role of the many variables which determine the final outcome of the various subgroups of patients.

Drug use in an intensive care unit and its relation to survival

Infection and trauma

Evaluation of fluid therapy in adult respiratory failure

Influence of total parenteral nutrition on fuel utilization in injury and sepsis

Nutrition for the patient with respiratory failure, glucose vs fat

Increased body temperature secondary to total parenteral nutrition

Milic-Emiii J. Nutrition and the respiratory system

Energy-requiring cell functions in the ischemic brain

Inappropriate ventilation and hypoxemia as causes of cardiac arrhythmias. The control of arrhythmias without antiarrhythmic drugs

Nutritional assessment: a comparison of clinical judgement and objective measurements

Stimulation of muscle protein degradation and prostaglandin E2 release by leukocyte pyrogen

Nitrogen-sparing effect of fat emulsion compared with glucose in the postoperative period

Inversed ratio ventilation (IRV)

Pharmacokinetics of highdose pentobarbital in severe head trauma

Pulmonary embolism: progress and problems

Current status of pulmonary thromboembolic disease: pathophysiology, diagnosis, prevention and treatment

Guidelines for drug therapy in renal failure

Massive blood transfusion

Nifedipine in hypertensive emergencies

Cerebral electrical activity influences the effects of etomidate on cerebral perfusion pressure in traumatic coma

Effects of major skeletal trauma on whole body protein turnover in man measured by L-I '4(: leucine

Branch chain amino acid administration and metabolism during starvation, injury and infection

Low-dose heparin prophylaxis of venous thrombosis: an editorial

Has clinical pharmacology lost its way? Lancet I

Hunt CEo The effect of independent variations in inspiratory-expiratory ratio and end expiratory pressure during mechanical ventilation in hyaline membrane disease: the significance of mean airway pressure

Pharmacokinetics of theobromine in male rats in different traumatic conditions

Baker WHo Comparison of hemodynamic, pulmonary, and renal effects of use of three types of fluids after major surgical procedures on the abdominal aorta

Phenytoin pharmacokinetics in burned rats and plasma protein binding of phenytoin in burned patients

Parenteral nutrition in surgical patients

Megadose steroids in severe head injury

Current application of "high -dose" steroid therapy for CNS injury

Diffuse cerebral swelling following head injuries in children: the syndrome of "malignant brain edema

Drug utilization in a general intensive care unit

Infection surveillance and control in the severely traumatized patient

Branched chains support to postoperative protein synthesis

The pharmacologic approach to the critically ill patient

Cardiovascular response to severe head injury

Muscle proteolysis induced by a circulating peptide in patients with sepsis or trauma

Effects of parenteral alimentation on amino acid metabolism in septic patients

Amino acid and energy metabolism in septic and traumatized patients

To dream the impossible dream

Cerebral autoregulation in unconscious patients with brain injury

Thrombolytic properties of human tissue-type plasminogen activator

Epidemiology of venous thromboembolism

Dexamethasone and severe head injury: a prospective double-blind study

Hemostasis in massively transfused trauma patients

Negative effect of albumin resuscitation for shock

Instability of lung units with low V A/Q ratios during 02 breathing

Nosocomial infections in intensive care wards: a multicenter prospective study

Plasma volume, intravascular protein content, hemodynamic and oxygen transport changes during intestinal shock in dogs

Oxygen toxicity of the lung: an update

Fat elimination in acute renal failure

Adult respiratory distress syndrome

Nosocomial infections -a 1980 view: progress, priorities and prognosis

Changes in nitrogen balance of depleted patients with increasing infusions of glucose

Branched chain amino-acids. What are they good for? Clinical Nutrition

A multicenter study of ICU drug utilization

Panel report on nutritional support of patients with liver, renal and cardiopulmonary disease

The effects of normalisation of plasma amino acid on hepatic encephalopathy in man

Plasma amino acids in patients with hepatic encephalopathy: effect of amino acid infusion

A phase I trial of naloxone treatment in acute spinal cord injury

Plasma amino acids as predictors of the severity and outcome of sepsis

Nitrogen sparing effects and mechanisms of branched -chain amino acids in the injured rat

Does intravenous fat spare nitrogen in the injured rat?

Apneic oxygenation in man

Treatment of acute respiratory failure with low-frequency positive-pressure ventilation and extracorporeal removal of CO2•

Bilancio energetico nel politraumatizzato durante alimentazione parenterale

A new look at therapy of the adult respiratory distress syndrome: motionless lungs

Handbook of Clinical Pharmacokinetics

Treatment of acute cerebral edema with high dose of dexamethasone

Nosocomial infections in a neonatal intensive care unit

The significance of colloid osmotic pressure measurement after crystalloid and colloid infusions

Failure of high -dose steroid therapy to influence intracranial pressure in patients with severe head injury

Colloid osmotic pressure and fluid resuscitation with hetastarch, albumin, and saline solutions

Oxygen transport responses to colloids and crystalloids in critically ill surgical patients

Controlled hyperventilation in patients with intracranial hypertension

Infections are a leading cause of mortality in a surgical intensive care unit

Treatment of cardiac and renal effects of PEEP with dopamine in patients with acute respiratory failure

Inotropic drugs in acute circulatory failure

Effect of branched-chain amino acid infusion on alterations in CSF neutral amino acid and their transport across the blood-brain barrier in hepatic encephalopathy

Prevention and treatment of brain ischaemia

Protein sparing and protein replacement in acutely injured patients during TPN with and without aminoacid supply

Nutritional support to long-term anesthetized and curarized patients under extracorporeal respiratory assistance for terminal pulmonary failure

Influence of total parenteral nutrition on protein metabolism following acute injury: assessment by urinary 3-methylhistidine excretion and nitrogen balance

Hyperammonemia plasma amino acid imbalance, and blood-brain amino acid transport: a unified theory of portal-systemic encephalopathy

Lipid infusion in critically ill patients: acute effects on hemodynamics and pulmonary gas exchange

Relationship of energy and protein input to nitrogen retention and substrate hormone profile

Studies in shock and resuscitation. I: use of a hypertonic, albumin-containing, fluid demand regimen in resuscitation

Continuous positive-pressure ventilation and portal flow in dogs with pulmonary edema

Altered coagulation after albumin supplements for treatment of oligemic shock

Enteral hyperalimentation

Alterations in cerebral blood flow, oxygen metabolism, and electrical activity produced by high dose sodium thiopental

Increased rates of whole body protein synthesis and breakdown in children recovering from bums

High level positive and expiratory pressure (PEEP) in acute respiratory insufficiency

Nutritional therapy for patients with acute renal failure

Pulmonary barotrauma during mechanical ventilation

Urinary excretion of sulfur amino acid and sulfur metabolites in burned patients receiving parenteral nutrition

Right and left ventricular geometry adjustments during acute respiratory failure

The management of acute renal failure

Elevated plasma binding cannot account for the bum-related d-tubocurarine hyposensitivity

Severe depression of antithrombin III associated with disseminated intravascular coagulation in women with fatty liver of pregnancy

Whole body protein synthesis and catabolism in septic man

Effect of carbohydrate and fat intake on nitrogen excretion during total intravenous feeding

Fresh blood by direct transfusion for haemostatic failure in massive haemorrhage

Crystalloid vs colloid in the etiology of pulmonary failure after trauma

Nutritional support of hospitalized patients

Impaired salt and water excretion after albumin resuscitation for hypovalemic shock

Impaired pulmonary function after albumin resuscitation from shock

A controlled clinical trial of streptokinase and heparin in the treatment of major pulmonary embolism

Glucose or fat as a non-protein energy source

Patterns of mortality in a surgical intensive care unit

The problems of blood transfusion

The critically ill or injured patient: priorities in management

Severe, acute disseminated intravascular coagulation

Influence of ventilation on response to fluid load in dogs: body water and albumin distribution

The outcome with aggressive treatment in severe head injuries. Part I

The outcome with aggressive treatment in severe head injuries. Part II

Hemodynamic and ventricular responses to pulmonary embolism

Multiple systems organ failure. I: the basal state

Multiple systems organ failure: II: the effect of infusion of amino acids and glucose

The surgical intensive care unit: current concept in infection

The techniques used to sedate ventilated patients

Intravenous infusion of ordinary and modified amino acid solutions in the management of hepatic encephalopathy

Head injury and brain ischaemia. Implications for therapy

Clinical infections, antibiotics and immunosuppression: a puzzling relationship

In vivo demonstration of the nitrogen sparing mechanisms for glucose and amino acids in the injured rat

Reduction in colloid osmotic pressure associated with fatal progression of cardiopulmonary failure

Intensive management of severe head injuries

Implications of nutrition in the surgical patient

Functional restoration of the traumatically injured spinal cord in cats by clonidine

Respiratory diseases task force report on problems, research, approaches and needs lung program of NHLI

Cerebral blood flow and metabolism in comatose patients with acute head injury

The influence of intravenous nutrition on protein dynamics following surgery

Digitalis and sympathomimetic stimulants

Cerebral circulation after head injury. IV. Functional anatomy and boundary-zone flow deprivation in the first week of traumatic coma

Pharmacokinetic and pharmacodynamic considerations in drug therapy of cardiac emergencies

Central dysregulation in brain stem lesions

Effect of megadose steroids on ICP in traumatic coma

Acute respiratory failure in the adult

Applied physiology in clinical respiratory care

Drug-related deaths among medical inpatients

Etude des facteurs de mortalite infectieuse dans un service de reanimation chirurgicale

Differential lung ventilation with PEEP in the treatment of unilateral pneumonia

Pulmonary morphology in a multihospital collaborative extracorporeal membrane oxygenation project

Disseminated intravascular coagulation. British Journal of Hospital Medicine

Pharmacologic management of acute intracranial hypertension

Care of the critically ill patient

The favorable effect of early parenteral feeding on survival in headinjured patients

High-dose barbiturate therapy in humans: a clinical review of 60 patients

Amelioration of post-ischemic brain damage with barbiturate

Prophylaxis of venous thromboembolism

Steroids in severe head injury: a prospective randomized clinical trial

Drug absorption and disposition in bum patients

Drug kinetics in bum patients

The effect of s0-lutions of varying branched-chain concentration on the plasma amino acid pattern and metabolism in intensive care patients

Positive end-expiratory pressure in acute lung injury

Intracranial hypertension

The use of fresh blood in the treatment of critically injured patients

Biochemical changes associated with severe trauma

Protein and caloric requirements with total parenteral nutrition

Total parenteral nutrition with Vamin and Intralipid

Pathophysiology and therapy of shock states

Fluid therapy in emergency resuscitation: clinical evaluation of colloid and crystalloid regimens

High rates and low volumes in mechanical ventilation -not just a matter of ventilatory fre-Quency

Pulmonary arterio-venous admixture. Improvement with albumin and diuretics

Alterations in plasma and CSF amino acids, amines and metabolites in hepatic coma

Optimal energy and nitrogen intake for gastroenterological patients requiring intravenous nutrition

Changes in protein synthesis after trauma: importance of nutrition

Pneumonia in an intensive care unit

Comparison by controlled clinical trial of streptokinase and heparin in treatment of life-threatening pulmonary embolism

Clinical relevance of pharmacokinetics

UPET-Urokinase Pulmonary Embolism Trial. A national cooperative study

USPET -Urokinase-Streptokinase Pulmonary Embolism Trial. Cooperative study

Crystalloids vs colloid resuscitation: is one better

Is i.v. administration of BCAA effective in the treatment of hepatic encephalopathy?

Colloid oncotic pressure: clinical significance

Cardiac output in bacterial shock

Positive end-expiratory pressure in adult respiratory failure

State of the art: intermittent mandatory ventilation

Acute respiratory failure

Insulin to inhibit protein catabolism after injury

Failure of prophylactically administered phenytoin to prevent early posttraumatic seizures

Laboratory of Clinical Pharmacology, Istituto di Ricerche Farmacologiche "Mario Negri