Critical Care Medicine
- 1. Description of the problem
2. Emergency management
- 3. Diagnosis
- 4. Specific treatment
- 5. Disease monitoring, follow-up and disposition
What's the evidence?
Also known as
Non-anion gap metabolic acidosis
Normal anion gap metabolic acidosis
Normal albumin-corrected anion gap acidosis
Normal strong ion gap metabolic acidosis
"Dilutional" or "infusion-related" acidosis
Post diabetic ketoacidosis
Renal tubular acidosis (RTA)
Urinary - enteric diversions
Other intestinal losses
Chronic renal disease
Acute renal failure
Carbonic anydrase inhibitors
Ammonium chloride administration
Total parenteral nutrition
1. Description of the problem
What every clinician needs to know
Hyperchloremic acidosis is a common acid-base disturbance in critical illness, often mild ( standard base excess >-10 mEq/L).
Definitions of hyperchloremic acidosis vary. The best are not based on chloride concentrations, but on the presence of metabolic acidosis plus the absence of significant concentrations of lactate or other unmeasured anions.
A useful definition is:
1. arterial pH less than 7.35,
2. standard base excess less than -3 mEq/L or bicarbonate less than 22 mmol/L,
3. Albumin corrected anion gap normal (5-15 mEq/L). A normal strong ion gap is an alternative indicator of the absence of unmeasured anions, although rarely used clinically and offering little advantage over the albumin corrected anion gap.
The degree of respiratory compensation is relevant. It is appropriate if PaCO2 approximates the two numbers after arterial pH decimal point (e.g. pH=7.25, PaCO2=25 mm Hg; this rule applies to any primary metabolic acidosis down to a pH of 7.1).
Acidosis is severe if standard base excess is less than -10 mEq/L, or pH is less than 7.3, or bicarbonate is less than 15 mmol/L.
Common causes in critical illness are large volume saline administration, large volume colloid infusions (e.g. unbalanced gelatine or starch preparations) following resolution of diabetic keto-acidosis or of other raised anion gap acidosis, and post hypocarbia.
Hyperchloremic acidosis often occurs on a background of renal impairment/tubular dysfunction. It is usually well tolerated, especially with appropriate respiratory compensation. The prognosis is largely that of the underlying condition. If associated with hyperkalemia, think of hypo-aldosteronism (Type 4 RTA), especially if diabetic. With persistent hypokalemia, think of RTA Types 1 and 2.
Clinical features of the condition
Hyperchloremic acidosis is usually well tolerated in the short term. Clinical features and associated mortality largely reflect the causal/associated condition(s).
Chronic metabolic acidosis (as occurs in RTA) can cause hypokalemia, hyperkalemia, phosphate and other electrolyte disturbances, bone mineral loss, muscle wasting, renal calculi and nephrocalcinosis.
Extreme acidemia (pH <7.0) is rare with non-anion gap metabolic acidosis. Even then, many adverse effects can be attributed to the underlying condition rather than the acidemia per se.
Pulmonary hypertension, respiratory muscle failure.
Myocardial depression, tachy and brady dysrhythmias, venoconstriction, vasodilatation with centralization of blood volume.
Increased metabolic rate, catabolism, ATP and 2,3-DPG depletion.
Cell membrane pump dysfunction.
Increased whole blood viscosity and hematocrit.
Hyperkalemia has been reported in renal transplant patients. Coagulation and platelet function may be affected (thromboelastography), with an increased blood product requirement in aortic aneurysm surgery.
Experimental hyperchloremia reduces renal and splanchnic blood flow, can cause nausea, vomiting and abdominal bloating, impairs performance of complex mental tasks, and has precipitated acute lung injury. There may be pro-inflammatory cytokine (IL-6) release and iNOS activation. Renal tubular excretion of free water may be impaired.
There are also potential benefits of metabolic acidosis. Lowering pH can protect against experimental hypoxic stress. Then there is the Bohr effect, in which reduced hemoglobin-oxygen affinity at low pH (rightward shift of the oxy-hemoglobin dissociation curve) enhances tissue oxygen unloading while maintaining unimpaired pulmonary capillary oxygen loading at normal ambient oxygen tensions. Over 24-48 hours the benefit of the Bohr effect is lost, since acidemia impairs phosphofructokinase activity, reducing erythrocytic 2,3-DPG production.
Key management points
Look for contributing factors and remove/correct if feasible.
IV sodium bicarbonate is rarely necessary, except in management of associated hyperkalemia. THAM is an alternative alkalinizing agent. In RTA types 1 and 2, regular oral alkalinizing therapy is often necessary.
2. Emergency management
Hyperchloremic acidosis can occur in the context of an emergency. It is rarely of itself an emergency. Emergency management is that of the underlying condition. IV sodium bicarbonate is rarely necessary, except in management of associated hyperkalemia. THAM is an alternative alkalinizing agent.
Emergency management of associated conditions where appropriate
If pH is less than 7.0 with cardiovascular compromise or hyperkalemia, consider IV sodium bicarbonate. In rare cases in which the disturbance includes hyperkalemia and renal failure, renal replacement therapy may be required.
Establishing a specific diagnosis
Laboratory diagnosis: Arterial pH is less than 7.35 (unless coupled with super-imposed respiratory alkalosis), standard base excess is less than 3 mEq/L or bicarbonate less than 22 mmol/L, albumin corrected anion gap less than 16 mEq/L.
Plasma [Cl] is usually, but not invariably elevated. [Cl] may be normal, or even low if there is hyponatremia accompanied by normal albumin concentrations. Remember that hyperchloremia without metabolic acidosis can also occur.
Normal lab values
Arterial pH less than 7.35 (unless the metabolic acidosis is coupled with an independent respiratory alkalosis), standard base excess less than -3 mEq/L or bicarbonate less than 22 mmol/L, albumin corrected anion gap less than 16 mEq/L. If these criteria are satisfied and the results are accurate, the patient has a 'hyperchloremic' type metabolic acidosis. The clinical context in which this can occur varies widely.
Plasma chloride 100-110 mmol/L
Plasma sodium 135-145 mmol/L
Plasma albumin 33-47 g/L
Arterial pH 7.35-7.45
PaCO235-45 mm Hg
Arterial plasma bicarbonate 22-27 mmol/L
Standard base excess -3 to +3 mEq/L
Anion gap 5-15 mEq/L
Albumin corrected anion gap 5-15 mEq/L
How do I know this is what the patient has?
Provided the criteria are satisfied and the measurements accurate, this is the predominant acid-base abnormality. To make this diagnosis it is not necessary for hyperchloremia to be present.
When the underlying cause of a non-anion gap metabolic acidosis is unclear, further investigations may be required. This is rarely necessary in intensive care practice.
Essentially the diagnostic sequence depends on the concentration of urinary ammonium, either de novo or after an ammonium chloride load, Urinary ammonium is reduced in RTA types 1 (distal) and 4, but present in appropriate concentrations in RTA type 2 (proximal) or with extra-renal causes of the acidosis, such as saline infusion or enteric losses.
Urine ammonium can be assayed formally via a 24 hour collection or its presence detected indirectly by calculating the urinary anion gap. A negative urinary anion gap indicates the presence of significant urinary ammonium concentrations.
The three possible causes are:
Administration of low SID fluids (e.g. 'dilutional' acidosis). (Should be self-evident)
Enteric losses of high SID fluid (diarrhea, pancreatic fistula etc.) or the presence of a urinary/enteric diversion. (Should be self-evident)
If options 1 and 2 seem unlikely, then Type 2 (proximal) RTA is a real possibility. It can be confirmed by demonstrating appropriate urinary acidification (pH <5.5) after ammonium chloride or furosemide administration, and that alkali loading causes an increased fractional bicarbonate excretion with a urine/blood PCO2gradient greater than 20 mm Hg.
Drugs and toxins that can cause this condition include acetazolamide and other carbonic anhydrase inhibitors, aminoglycosides, valproate, chemotherapeutic agents and heavy metals. Phosphaturia and other proximal tubular losses occur in Fanconi syndrome. Other causes include light chain nephropathy, amyloidosis and paroxysmal nocturnal hemoglobinuria.
The plasma potassium concentration distinguishes the two main possible causes here:
Raised plasma potassium supports the diagnosis of Type 4 RTA. Urine pH will be less than 5.5 after an acid load. (If urinary pH >5.5, the diagnosis is more likely a hyperkalemic variant of distal RTA). Further work up then includes plasma renin and aldosterone concentrations (to diagnose mineralocorticoid deficiency or resistance), plasma free cortisol before and after synthetic ACTH (to detect hypo-adrenalism), and the investigation of a possible underlying nephropathy. Examples of drugs that may cause Type 4 RTA include ACE inhibitors, heparin, potassium retaining diuretics and beta blockers.
Normal or low plasma potassium. The diagnosis is most likely to be Type 1 (distal) RTA. In this case ammonium chloride loading or frusemide administration will fail to acidify urine pH below 5.5. Supportive features include a urine/blood PCO2gradient less than 20 mm Hg after alkali loading or frusemide. Many inherited and acquired conditions can cause distal RTA, including rheumatoid arthritis, systemic lupus erythematosis, primary biliary cirrhosis, renal transplant rejection, post-obstructive uropathy and primary hyperparathyroidism. Drugs include amphotericin B and lithium carbonate.
4. Specific treatment
With removal of the underlying cause and provided there is adequate renal function, the acidosis should resolve in 24 to 48 hours. If further volume loading is required, it should be with a balanced fluid such as compound sodium (Ringer's) lactate solution rather than saline.
On the other hand, if there is volume overload rather than hypovolemia, intravenous furosemide will accelerate resolution of the metabolic acidosis (by causing a diuresis in which the urine has a reduced strong ion difference due to inhibition of chloride resorption).
Meanwhile, if the patient is receiving mechanical ventilation, aim for a minute volume which provides appropriate respiratory compensation. This may not be feasible if restricted minute volumes are necessary due to ARDS or acute lung injury, in which case intravenous sodium bicarbonate administered slowly will lessen the severity of the acidemia. THAM is an alternative alkalinizing agent.
For obese patients, use approximate ideal body weight rather than actual body weight or dosing body weight. Side effects include hyper-osmolarity, hypokalemia, ionized hypocalcemia and a sudden increase in hemoglobin-oxygen affinity.
For obese patients, use approximate ideal body weight rather than actual body weight or dosing body weight. Side effects include apnea (due to sudden CNS hypocarbia), hypoglycemia, dyskalemias and coagulation disturbances. THAM is renally excreted and accumulates in renal dysfunction with repeated dosing.
5. Disease monitoring, follow-up and disposition
Expected response to treatment
With cause removal, expect normalization of pH, plasma bicarbonate and standard base excess over 24 to 48 hours provided there is adequate renal function. Expect more rapid resolution with frusemide therapy. There will be an immediate dose-related response with sodium bicarbonate or THAM administration.
The prognosis is that of the underlying condition. The risk of an adverse outcome due to the acid-base disturbance itself is small.
There may be hyperlactemia (>3 mmol/L), but as a minor component of the acidosis. Other non-chloride anions such as the ketone bodies aceto-acetate and beta-hydroxyacetate may also be present. Without obvious elevation of the albumin corrected anion gap, the contribution of non-chloride anions to the metabolic acidosis should be small.
Various 'improvements' on the anion gap as a scanning tool for unmeasured anions have been suggested. These include the albumin corrected anion gap, the base excess gap, the strong ion gap and the 'net unmeasured anion' concentration. In one (currently unpublished) study comparing detection of unmeasured anions by various scanning tools, the albumin corrected anion gap had the highest area under the receiver operator characteristic curve when compared with the anion gap and the base excess gap (0.78 versus 0.56 versus 0.62 respectively). In the same study the strong ion gap, which because of its complexity is less clinically convenient, performed no better than the albumin corrected anion gap (ROC area 0.78).
The chloride ion-selective electrode is prone to variation and interferences. For example, bromism and hyperlipidemia can cause chloride overestimation and a falsely normal albumin corrected anion gap. Confirmation of chloride concentrations on two instruments (laboratory and point of care) is helpful if there is doubt.
With the advent of modern ion selective electrodes, there has been an upshift in the normal chloride reference range, although this has varied from manufacturer to manufacturer. As a result there has been a corresponding downshift in reference values for the anion gap, corrected anion gap and gap scanning tools in general. Hence it is vital that laboratories regularly calibrate these measured and derived parameters against the local reference population.
Another cause of a 'falsely normal' albumin corrected anion gap or strong ion gap is the presence of high concentrations of unmeasured cations. This can occur in lithium overdose, IgG myeloma, or after THAM administration. Severe hypernatremia can cause sodium and thus anion gap under-estimation. Severe hyper-albuminemia causes sodium under-estimation only when using indirect ion-selective electrodes.
Follow up with at least 2 further blood gas and electrolyte analyses over the next 24 hrs, or until the condition resolves.
The easiest way to understand so-called 'hyperchloremic metabolic acidosis' is via the Stewart "physical chemical" approach to acid-base analysis. At its simplest, the mechanism of this disturbance can be thought of as follows:
The plasma chloride concentration on its own does
By the Principle of Electroneutrality, a narrowed [Na]-[Cl] concentration difference creates pre-conditions for metabolic acidosis by reducing the negative charge 'space' available for the bicarbonate anion. When the sodium concentration is normal, a significant reduction in the [Na]-[Cl] difference must cause hyperchloremia, in keeping with the classic 'hyperchloremic acidosis' concept. However, if there is hyponatremia, a 'hyperchloremic' type metabolic acidosis may be present despite a normal or even a low chloride concentration.
A separate determinant of metabolic acid-base status is the 'non-CO2' (non-volatile) weak acid concentration in plasma. This is primarily due to albumin, with a smaller contribution from inorganic phosphate. Both molecules exhibit weak acid activity.
Reduced weak acid activity (hypo-albuminemia) in isolation causes a metabolic alkalosis. The only way to counteract the metabolic alkalosis of hypo-albuminemia is via an accompanying reduction in the [Na]-[Cl] difference. In this circumstance we will have
Hence in 'hyperchloremic' type acidosis, we always find a [Na]-[Cl] difference that is low (except in the rare situation where the albumin concentration is elevated). However if the sodium concentration is also low,
Causes can be divided into two broad categories:
1. Loss of large volumes of 'high [Na]-[Cl] difference' fluid - for example in RTA (urine) or in some cases of diarrhea (enteric contents).
2. Gain of large volumes of low [Na]-[Cl] difference fluid. The example here is metabolic acidosis due to saline infusion, where the [Na]-[Cl] difference of the infused fluid is zero. This type of abnormality has been termed "dilutional acidosis." The same phenomenon can occur with fluids of varying chloride content (including 0.45% saline, dextrose saline combinations and colloids). In each case the fluid [Na]-[Cl] difference is either zero, or else low enough to decrease the plasma [Na]-[Cl] difference at a rate that overwhelms the concurrent dilutional reduction in albumin and phosphate, which would otherwise cause a metabolic alkalosis.
In either of these scenarios, biochemical scanning tools such as the anion gap, albumin-corrected anion gap or strong ion gap will not be increased. This means that anions like keto-acids, salicylate, glycolate and others are unlikely to be present in sufficient concentrations to cause a metabolic acidosis on their own.
More detail on the Stewart approach to acid-base as applied to metabolic acidosis:
In the Stewart paradigm, metabolic acid-base status is a function of two independent variables interacting in intravascular and interstitial compartments. These are the strong ion difference (SID) and the total concentration of non-volatile weak acid (ATOT). SID is the net charge in mEq/L of all fully dissociated ions, such as sodium potassium, calcium, magnesium, chloride, lactate and the keto-anions. The plasma SID is normally around 42 mEq/L. ATOT=[HA]+[A-], where HA denotes a non-volatile weak acid in equilibrium with dissociation products A- and H+.
Extracellular ATOT consists of albumin and phosphate, Intra-erythrocytic A TOT, primarily hemoglobin, also plays an important part in any final acid-base equilibrium. PCO2, the third and final independent variable, determines respiratory acid-base status. All three independent variables (SID, ATOTand PCO2) act in concert to determine the fluid pH as well as the values of other dependent variables such as [HCO3-]. From the stand-point of metabolic acid-base status, an isolated increase in ATOT or decrease in SID creates a metabolic acidosis, while changes in the opposite directions respectively cause a metabolic alkalosis.
Hence from the perspective of physical chemistry, plasma [Cl-] should not be considered in isolation when assessing the mechanism of a metabolic acid-base disturbance, since it is only one of several strong ions affecting the SID. Its value, together with concentrations of other strong anions, is relevant only in conjunction with the accompanying strong cations, especially [Na+], the principle strong cation. Metabolic acidosis means extracellular SID is low when matched against the prevailing ATOT.
As a general rule, non-anion gap acidosis can arise in two ways. In both cases renal acid-base homeostasis, which normally acts to restore an appropriate extracellular SID by altering urinary SID, is either overwhelmed due to the rapidity of the process or is itself malfunctioning. The two mechanisms are:
1. Excessive loss of high SID fluid
2. Excessive gain of low SID fluid
However, fluid induced metabolic acidosis can also result from infusions containing low [Cl -] such as 0.45% saline, or zero [Cl-] such as mannitol. The relevant crystalloid property is not [Cl-] alone, but its SID. Extracellular SID falls at the same rate in response to any zero SID infusion, whether the fluid administered has a low, normal or high [Cl-]. With low [Cl-] infusions, this will be accompanied by an unchanged or falling extracellular [Cl-], but always with a greater reduction in [Na +].
With colloid preparations, the situation can be more complex. As with crystalloids, the final result is determined by the equilibratium extracellular SID and ATOTafter being forced in the direction of the SID and the ATOT of the infused fluid. Albumin and gelatin are weak acids. In other words, from a Stewart perspective they qualify as ATOT. However these preparations are also pH adjusted with NaOH, which raises their SID above zero.
The net result, at least in vitro, is an identical tendency to cause metabolic acidosis on infusion to that of saline, although hyperchloremia is less prominent and there is no dilutional effect on ATOT. On the other hand, starches and dextrans have no weak acid activity. This means that their acid-base effects are determined by their excipients (usually saline).
Renal tubular NH3+ acts as a variable cationic partner for tubular Cl- and for other urinary strong anions, particularly sulphate and hippurate, which are produced constantly (50 mEq/day) as end-products of protein metabolism. NH4+ up or down-regulation allows for an adjustable urinary SID, by substituting in tubular electroneutrality transactions for an equal concentration of Na+ .
In renal tubular acidosis, the urinary SID 'setting' is inappropriately high, and in some variants there is a shallow urinary SID nadir following an acid load. In Types 1 and 4 RTA, the problem is insufficient upregulation of urinary NH3+, and in Type 2 there is excessive proximal tubular resorption of urinary Cl-.
Hyperchloremia is common in critical illness. It has been reported in up to 80% of patients in a mixed medical-surgical ICU. 'Severe' hyperchloremia ([Cl-] > 114 mol/L) occurs less frequently (around 6% in one recent report), and the prevalence of metabolic acidosis of any kind, hyperchloremic or otherwise, is also lower. However, the lack of a uniform definition has been a major problem, particularly in reports from the pre-Stewart era.
Estimates of incidence or prevalence of all acid-base disturbances thus vary greatly, depending on definitions as well as the case-mix in question. Even in recent reports where Stewart style criteria have been applied, estimates of the incidence of 'hyperchloremic' type acidosis in critically ill populations range from less than 10% to more than 60%.
There is now evidence that restricting the use of 'chloride rich' fluids in ICU can reduce the incidence of hyperchloremia, metabolic acidosis and acidemia, while increasing the incidence of metabolic alkalosis and alkalemia. It remains to be established whether altering practice in this way produces any effect on important measurable outcomes such as the occurrence of renal failure, the time requiring ventilatory support, the length of ICU stay or mortality.
As with the data on epidemiology, a major problem in determining the prognosis of non-anion gap acidosis is the lack of a uniform definition. The best published estimate in a group of critically ill patients using a valid physical chemical definition puts the overall mortality at 30%. Of note, the reported mortality of conditions associated with hyperlactemia or with a raised strong ion gap acidosis is generally higher at 40-60%.
In reality the prognosis of non-anion gap acidosis is largely that of the underlying condition, rather than of the acid-base disturbance itself. For example if hyperchloremic acidosis occurs in the context of fluid resuscitation for a ruptured abdominal aortic aneurysm, a mortality rate of at least 30% can be expected. However, a non-anion gap acidosis invariably appears after resuscitation of diabetic ketoacidosis. On paper the post-DKA disturbance is often moderately severe (standard base excess < -10 mEq/L), yet this has little or no association with morbidity or indeed mortality.
However, long standing metabolic acidosis, as occurs in the various types of RTA, does carry significant morbidity, for example hypokalemia, hyperkalemia, phosphate and other electrolyte disturbances, bone mineral loss, muscle wasting, renal calculi and nephrocalcinosis.
What's the evidence?
Morgan, TJ, Bersten, AD, Soni, N. "Acid-base balance and disorders. In: Oh’s Intensive Care Manual". Butterworth-Heinemann Elsevier. 2009. pp. 949-61.(This book chapter has relevance to most sections.)
Handy, JM, Soni, N. "Physiological effects of hyperchloraemia and acidosis". Br J Anaesth. vol. 101. 2008. pp. 141-50.(This article is an important source concerning the clinical features of hyperchloremic acidosis.)
Soriano, JR. "Renal tubular acidosis; The clinical entity". J Am Soc Nephrol. vol. 13. 2002. pp. 2160-170.
Gluck, SL. "Acid-base". Lancet. vol. 352. 1998. pp. 474-9.(Articles 3 and 4 above are bicarbonate based in their approach, whereas the author's preference is the physical chemical approach. Nevertheless, they are useful sources of information concerning the classification, diagnosis and management of renal tubular acidosis in particular.)
Morgan, TJ, Kellum, JA, Elbers, P.W.G.. "Unmeasured Ions and the Strong ion Gap". Stewart's Textbook of Acid Base. AcidBase.org. 2009. pp. 323-37.(This book chapter contains extensive descriptions and analyses of the strengths and weaknesses of the various scanning tools for unmeasured anions.)
Morgan, TJ. "The meaning of acid-base abnormalities in the intensive care unit: part III - effects of fluid administration". Crit Care. vol. 9. 2005. pp. 204-11.
Morgan, TJ, Ronco, C, Bellomo, R., Kellum, J.A.. "Iatrogenic Hyperchloremic Metabolic Acidosis". Critical Care Nephrology. Saunders Elsevier. 2009. pp. 651-5.
Morgan, TJ, Kellum, JA, Elbers, P.W.G.. "Fluid Resuscitation". Stewart's Textbook of Acid Base. AcidBase.org. 2009. pp. 351-63.(References 6-8 above deal in detail with the etiology and pathophysiology of fluid induced hyperchloremic acidosis ['dilutional' acidosis].)
Gunnerson, KJ, Saul, M, Kellum, JA. "Lactate versus non-lactate metabolic acidosis: a retrospective outcome evaluation of critically ill patients". Crit Care. vol. 10. 2006. pp. R22.
Gunnerson, KJ. "Clinical review: the meaning of acid-base abnormalities in the intensive care unit part 1 - epidemiology". Crit Care. vol. 9. 2005. pp. 508-16.(References 9 and 10 written from the physical chemical perspective, are source material for information on epidemiology and prognosis.)
Copyright © 2017, 2013 Decision Support in Medicine, LLC. All rights reserved.
No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.
Sign Up for Free e-newsletters
- New Hypertension Threshold Guides Blood Pressure Management During Cancer Treatment
- Sexual Aids and Resources Not Readily Available at Cancer Centers
- Exercise Before Lung Cancer Surgery Greatly Reduces Complications
- 5-Year Overall Survival in Endometrial Cancer Not Improved With Chemoradiotherapy
- Factors Affecting Employment Participation in Early-stage Breast Cancer
- Sitting With Silence in End-of-Life Cancer Care
- Obesity and Cancer Risk (Fact Sheet)
- Susceptibility Gene Mutations Common in Those With Pancreatic Cancer and History of Other Cancers
- Anticancer Properties of Omega-3 Fatty Acids: Plant-Based vs Marine-Based
- US Pharmacopeia Revises Chapter on Handling Hazardous Drugs
- Follow-Up After Treatment Lacking for Younger Cancer Survivors
- Adjuvant Vemurafenib Does Not Improve Disease-Free Survival in Melanoma
- Profile of Buparlisib and its Potential in the Treatment of Breast Cancer: Evidence to Date
- Exercise Before Lung Cancer Surgery Greatly Reduces Complications
- Hodgkin Lymphoma Treatment in EU vs US: Similarities Would Enable Worldwide Studies
Regimen and Drug Listings
GET FULL LISTINGS OF TREATMENT Regimens and Drug INFORMATION
|Head and Neck Cancer||Regimens||Drugs|